JP4941316B2 - Image reading device - Google Patents

Description

The present invention relates to an image reading apparatus having a to that function analyzes the image data representing the read image of the originals.

  2. Description of the Related Art Conventionally, as an image reading apparatus, a flat bed type image reading apparatus that includes a reading unit under a reading glass and reads a document placed on the reading glass through the reading unit under the reading glass is known. In this type of image reading apparatus, while the reading unit is conveyed under the reading glass, the reading unit is caused to perform a reading operation during the conveyance, thereby reading a document placed on the reading glass and representing a read image of the document. Generate image data.

  In addition, an image reading apparatus having a copy function is known. As an image reading apparatus having a copy function, a digital multifunction peripheral having a printer function, a fax (FAX) function, etc. in addition to a copy function. It has been known.

Also, as an image reading apparatus, an apparatus that reads a document to be read through a reading unit and detects a document edge (edge of the document) from the reading result is known (for example, Patent Documents 1 to 5). 3). An image reading apparatus having a copy function detects the size and inclination of a document to be read from a reading result obtained by pre-scanning (see, for example, Patent Document 4), and uses the detected document size and inclination information. Based on this, an image reading apparatus that sets a variable magnification, an inclination correction amount, and the like is known.
JP 2004-201240 A JP 2005-285010 A JP 2005-039620 A JP 2000-022898 A

  By the way, in the method of detecting the document edge that is the edge of the document from the result of reading the document by the reading unit, there is a possibility that the ruled line drawn in the document is erroneously detected as the document edge. That is, in the image data representing the read image of the document, an image pattern similar to the edge of the document appears even at a point where the ruled line is present in the document. On the other hand, since the original to be read is tilted or the original size is indeterminate, an image pattern corresponding to the edge of the original is found from the above image data as a read result and the original edge is detected. However, there is a possibility that an image pattern such as a ruled line inside a document is erroneously detected and a document edge is erroneously detected.

  Further, in a flat bed type image reading apparatus, dust (hair, etc.) is placed on the reading glass in order to read an image on the reading glass under the reading glass with a document placed on the reading glass. Then, the dust is also read, and an image other than the original appears in the image data as the read result. Therefore, even when image data in which such a dust image appears is analyzed to detect a document edge, the document edge may be misled by the dust image.

As described above, in the method of detecting the document edge from the result of reading the document by the reading unit, the document edge may be erroneously detected due to various factors.
The present invention has been made in view of these problems, and suppresses the influence of ruled lines, dust, and the like inside a document in an apparatus that detects a document edge (document edge) from image data obtained as a result of reading by a reading unit. An object of the present invention is to provide a technique capable of appropriately detecting an edge of a document.

Is made the inventions in order to achieve the above object, a document table document to be read is placed, a reading unit for reading a document placed on the document table, the reading unit, mounting the platen by reading the location to the document, an image reading apparatus Ru comprising a reading control means for generating image data representing the read image of a document placed on the document table, the further conversion means, selection Means, storage means, and detection means.

The conversion unit in the image reading apparatus converts the image data generated by the reading control unit into image data representing an edge image. Further, the selection means selects, from the image data converted by the conversion means , candidate edge points representing the edge of the document extending in a specific direction on the document table . The storage means stores position information of the edge point selected as a candidate by the selection means.

In addition, the detecting means is selected from a group of candidates selected by the selecting means based on the distribution of edge points selected as candidates by the selecting means specified from the position information of the edge points stored in the storing means. By discarding edge points with low accuracy as edge points representing the edge of the document, candidates remaining after the discard are detected as edge points representing the edge of the document extending in the specific direction.

  In addition, the selection unit specifically sets the end point of the converted image data as a reference start position, and extends from the reference start position to the inner side of the image data along a predetermined first direction. Then, by referring to the pixel value of each pixel constituting the image data, the edge point to be inspected for continuity is detected, and the detected edge point is the starting point and is perpendicular to the first direction. A predetermined amount of edges that are continuous in the second direction by checking the continuity of the edge points in the second direction with reference to the pixel values of the pixels that make up this image data in the second direction A point is selected from the image data as a candidate for an edge point representing the edge of the document extending in a specific direction.

  The selection unit repeatedly performs such an operation a plurality of times while shifting the reference start position in the second direction, thereby generating a predetermined amount of edge point candidates representing the edge of the document extending in a specific direction. Search and select from the end of the data.

As described above, in the image reading apparatus according to the present invention, the plurality of edge point candidates representing the edge of the document are searched and selected from the edge of the image data in units of the predetermined amount. The edge point (document edge) representing the edge of the document can be accurately detected while suppressing the influence of the dust and the ruled lines drawn inside the document.

  That is, according to the present invention, the candidate of the edge point representing the edge of the document is independently searched and selected by a predetermined amount, so that some of the selection operations that are performed multiple times are erroneously selected. Even in the corresponding operation, the influence of the erroneous selection does not affect the selection operation of other times, and it is possible to suppress the erroneous selection from spreading to the entire selection operation of a plurality of times.

Therefore, according to this image reading apparatus , even when an erroneous edge point is selected as an edge point candidate representing the edge of the document, the erroneous selection operation is corrected by the operation of the detection means. Finally, the correct document edge (edge point representing the edge of the document) can be detected, the influence of ruled lines and dust inside the document can be suppressed, and the document edge can be detected appropriately.

By the way, the reading control means can be configured to have a function of causing the reading unit to read an area shorter than the area extending from one side of the document table perpendicular to the second direction to the center extending in the second direction. The selection means can be configured to select edge point candidates from the image data converted by the conversion means for the image data generated by the reading control means by this function. In addition, the detection unit sets a predetermined amount of candidates selected by the selection unit as one group unit, and the continuity of each group in the second direction is determined based on the group closest to the center of the document table in the second direction. evaluated, all candidate belonging to discontinuous groups and groups and the reference, Ru can certainty is to destroy configured as low candidates.

In the image reading apparatus described above, the direction in which the edge of the document to be detected extends when the document is correctly placed on the document table is defined as the second direction, and the direction perpendicular thereto is defined as the first direction. In this way, the target document edge can be detected. However, when the first and second directions are determined in this way, the document is not placed correctly on the document table, but placed at an angle. The edge point corresponding to the document edge different from the edge of the document to be detected may be detected as the edge point to be inspected for continuity. There is a possibility that a ruled line or the like drawn inside is erroneously selected as a candidate for an edge point representing the edge of the document (see FIG. 9A).

  Accordingly, it is preferable that the erroneously selected edge point can be discarded by the detecting means. However, an erroneous selection operation that occurs due to the inclination of the document may occur on the end side of the document table. Is high and is unlikely to occur at the center of the platen.

  For this reason, the detection means described above has a high probability that the edge point located near the center of the document table among the edge points selected as candidates by the selection means is an edge point representing the edge of the document. It is treated as an edge point, and all of the candidates belonging to a group discontinuous from the group closest to the center of the document table in the second direction are discarded as candidates with low accuracy.

Therefore, according to the inventions of an image reading apparatus including the detecting means, an erroneous selection operation of the selection means according to the document is tilted, with modification detection means able to detect the correct document edge, Document edges can be detected appropriately without being affected by the ruled lines of the document.

Further, the detection means may comprise an inspection object switching setting means, a first determination unit, a divergence degree calculating unit, a second determination means.

  The inspection target switching setting means sets the inspection target group so as to be sequentially switched from the group closest to the center of the document table to the endmost group in the second direction, and the first determination means switches the inspection target switching. In the direction in which the group to be inspected is switched by the setting means, a one-dimensional parallel to the second direction indicates whether or not an adjacent group exists in the group to be inspected currently set by the inspection target switching setting means. Evaluate and judge in space. That is, it is determined by evaluating in the second direction whether or not a group adjacent to the group to be inspected exists in a direction in which the group to be inspected is switched by the inspection target switching setting unit. At this time, the first direction is not evaluated.

  When the first determination unit determines that there is an adjacent group, the divergence degree calculation unit is a candidate belonging to the group to be inspected and closest to the group, and a candidate belonging to the adjacent group Then, the degree of deviation in the first direction from the candidate closest to the group to be examined is calculated (see FIG. 9B). Further, the second determination unit determines whether or not the divergence degree calculated by the divergence degree calculation unit is less than a threshold value.

  When the first determination unit determines that the adjacent group does not exist or the second determination unit determines that the divergence degree is equal to or greater than the threshold, the detection unit determines that the group to be inspected by that time point. All of the groups that were not set (groups located on the edge side of the platen in the second direction) are considered as discontinuous groups from the group closest to the center of the platen in the second direction. Discard all candidates belonging to the discontinuous group.

By configuring the detection means in this way, if all of the candidates belonging to a group discontinuous from the group closest to the center of the document table in the second direction are discarded, the same as in the above-described invention . An effect can be obtained.

The detection hand stage, as a group unit candidates selected predetermined amount by the selection means, a set of groups to evaluate the continuity of the second direction of each group, successive longest probable A candidate may be regarded as a set of candidates, and each candidate belonging to a group that does not belong to the longest continuous group set may be regarded as a candidate with low accuracy, and a candidate with low accuracy may be discarded.

  The set of edge points that continue long in the second direction is more likely to be the edge point that truly represents the edge of the original document to be detected. There is a high possibility that the edge point is caused by dust or the like placed on the document table.

  Therefore, if the detection means is configured in this way so that a set of edge points that are longest in the second direction is detected as an edge point that represents the edge of the document, dust or the like placed on the document table The set of edge points erroneously selected by the selection means due to the influence of the selection means can be discarded by the detection means, and the document edge can be detected correctly, and the influence of dust etc. can be suppressed and the document edge can be detected appropriately. .

Further, the detection means includes an inspection object switching setting means for setting the inspection object group so as to sequentially switch in the second direction, the first determination means, the divergence degree calculation means, and the second determination means. For the group to be inspected that is determined by the first determination means that there is an adjacent group and the degree of divergence is determined to be less than the threshold by the second determination means. It is evaluated that the adjacent group set as the group to be inspected is continuous, and it is determined by the first determination unit that there is no adjacent group, or the second determination unit determines that the divergence degree is equal to or greater than a threshold value For a group to be inspected that is determined to be, the group and the group set as the group to be inspected next to the group are evaluated as not consecutive. , It can be configured to detect a set of successive groups the longest.

In addition, the detecting means, a candidate of a predetermined amount selected by selecting means as a group unit, calculates the approximate line of each group, based on the calculated slope and intercept information of the approximate straight line in each group, Generate histogram data for the approximate line, and based on the generated histogram data, consider the set of groups for which the approximate line with the highest frequency is calculated as a candidate set with the highest accuracy, and calculate the approximate line with the highest frequency By detecting each candidate belonging to a group that does not belong to the group of the selected group as a low-accuracy candidate and discarding the low-accuracy candidate, an edge point representing the edge of the document extending in the specific direction is detected. Can be configured.

  In other words, the detecting means can detect a set of groups corresponding to the approximate straight line having the highest frequency among the approximate straight lines of each group as a set of edge points representing the edge of the document.

  If each group is a set of edge points that truly represent the edge of the document, it should be understood that the slope and intercept of the approximate line should match. In other words, the slope of the approximate line of each group and the similarity of the intercepts are evaluated by a histogram, and a set of edge points corresponding to the approximate line with the highest frequency is detected as an edge point representing the edge of the document. Thus, the document edge can be detected with high accuracy.

Thus, this technique, select the candidate by a predetermined amount in the selecting means, by detecting the edge points representing the edge of the document by the detection means, the effects of the placed dust and document internal borders such as platen The edge of the document can be detected correctly with as little as possible.

Also, edge point information (document edge information) representing the edge of the document detected by the image reading apparatus can be used for, for example, OCR processing, a variable magnification / tilt correction amount setting operation, and the like. The

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1A is a block diagram illustrating the configuration of the digital multifunction peripheral 1 according to the present embodiment. As shown in FIG. 1A, the digital multifunction peripheral 1 of this embodiment includes an image reading unit 10, a reading control unit 20, a printing unit 30, a printing control unit 40, a display operation unit 50, a communication unit 60, a CPU 70, A RAM 80 and a flash memory 90 are provided, and various processes are executed by the CPU 70 based on a program recorded in the flash memory 90 to control the entire apparatus.

  Specifically, the image reading unit 10 is configured as shown in FIGS. FIGS. 1B and 1C are explanatory diagrams showing the configuration of the image reading unit 10. The image reading unit 10 is configured such that a transparent plate-shaped reading glass 11 on which a document P to be read is placed is supported by a housing 13, and the reading is performed below the reading glass 11 in the housing 13. A reading unit 15 that optically reads an image reflected on the glass 11 in the main scanning direction, a conveying mechanism 16 that conveys the reading unit 15 in the sub-scanning direction, and a motor 17 that drives the conveying mechanism 16 are provided.

  The image reading unit 10 drives the transport mechanism 16 by the rotational force of the motor 17 in accordance with a control signal input from the reading control unit 20, thereby causing the reading unit 15 to move under the reading glass 11 in the sub-scanning direction (FIG. 1 ( b) While being conveyed in the direction of the dotted arrow shown in (c), the reading unit 15 is caused to perform a reading operation during conveyance.

  Note that the reading unit 15 is controlled by the reading control unit 20 and, as it moves in the sub-scanning direction, reads an image shown on the reading glass 11 line by line and outputs a line image signal representing the read image. (CIS). The line image signal output from the reading unit 15 is converted into digital data (line image data) by an A / D converter (not shown) and input to the reading control unit 20.

  The image reading unit 10 is provided with a lid (not shown) that can cover the reading glass 11 so as to be openable and closable similarly to a known scanner. When the reading operation by the reading unit 15 is performed, a manual operation by the user is performed. By the operation, the lid is closed so as to be placed on the document placed on the reading glass 11. A white member is provided on the inner side of the lid facing the reading glass 11 so that a black shadow does not appear in the outer area of the document in the image data that is the reading result of the reading unit 15.

  On the other hand, the reading control unit 20 executes a reading control process in accordance with an instruction from the CPU 70, and controls the movement of the reading unit 15 in the sub-scanning direction and controls the reading operation of the reading unit 15 in the reading control process. To do.

  With this control, the reading control unit 20 causes the reading unit 15 to read an image reflected in the reading area of the reading glass 11 designated by the CPU 70 while conveying the reading unit 15 under the reading glass 11 in the sub-scanning direction. Image data representing an image shown in the reading area of the reading glass is recorded in the RAM 80. By such an operation, image data (two-dimensional image data in which pixels are arranged in the main scanning direction and the sub scanning direction) representing the read image of the document P placed in the reading area of the reading glass 11 is stored in the RAM 80. To be recorded.

  The reading control unit 20 temporarily stores each line image data input from the image reading unit 10 in a built-in buffer, performs image processing such as shading correction on each line image data, and then executes each line image data. Image data is recorded in the RAM 80.

  In addition, the printing unit 30 conveys a recording sheet placed on a tray (not shown) to a recording position in accordance with a control signal input from the printing control unit 40, and is well-known such as an ink jet method or a laser printer method. In this recording method, an image corresponding to the control signal is formed on the recording paper.

  The print control unit 40 controls the printing unit 30 to form an image on recording paper. In accordance with a command from the CPU 70, an image based on the print target data designated by the CPU 70 is transmitted through the printing unit 30. Print on recording paper.

  The communication unit 60 includes a group of interfaces for communicating with an external device, and includes a USB interface, a LAN interface, a FAX modem, and the like. That is, the multi-function device 1 is configured to be capable of performing FAX communication with an external FAX device through a FAX modem provided in the communication unit 60, and configured to be capable of communicating with an external personal computer through a USB interface or a LAN interface.

  In addition, the display operation unit 50 includes a liquid crystal display (not shown) for displaying information and various operation keys, and is controlled by the CPU 70 to display information for the user on the liquid crystal display. The command from the user input through is input to the CPU 70.

  Further, the CPU 70 realizes a copy function, a FAX function, a scanner function, a printer function, and the like according to a command input through an operation key or a command input from an external personal computer or the like through the communication unit 60 by executing a program. To do.

  For example, when a copy command is input from the user through an operation key provided on the display operation unit 50, the CPU 70 executes a copy control process (see FIG. 3) to control each unit in the apparatus, and on the reading glass 11. The read image of the placed document P is printed on a recording sheet. More specifically, the size of the document P placed on the reading glass 11 is estimated based on the reading result by the reading unit 15, and the scaling ratio is set from the estimated document size and the recording paper size, and the document P The read image is enlarged or reduced to a size suitable for the recording paper size, and this is printed on the recording paper through the printing unit 30 (automatic scaling copying process).

The contents of the copy control process will be specifically described below. Prior to that, the characteristics of the image reading unit 10 of this embodiment will be described with reference to FIG.
FIG. 2 is an explanatory view showing the readable region R0 of the reading glass 11. As shown in FIG. In the multifunction device 1, a reading glass 11 that is a rectangular glass plate is provided so as to close the opening portion in the opening portion of the rectangular housing 13 whose one surface is opened. The periphery of the reading glass 11 is supported by a housing 13.

  The reading glass 11 is provided slightly below the upper surface of the housing 13 from the upper surface of the housing 13, and an area 11 a of the reading glass 11 exposed from the housing 13 (hereinafter, referred to as “original table”). A step is formed at the boundary BD between the casing 13 and the casing 13 so that the original P can be abutted (hereinafter referred to as a portion 13a above the reading glass 11 of the casing 13 (the portion forming the step)). Expressed as “frame”.)

  Further, in the image reading unit 10, due to the relationship between the line width of the reading unit 15 and the size of the document table 11 a, a rectangular region R 0 that is slightly narrower than the entire region of the rectangular document table 11 a is the reading unit 15. Is defined as a readable area R0 where the original can be read. Specifically, the readable area R0 is an area having an outer periphery at a position away from the boundary BD between the frame 13a and the reading glass 11 by a minute amount (3 mm in this embodiment), as indicated by a dotted line in FIG. ing.

  In the image reading unit 10, a mark MK is instructed in the lower left part of the frame 13a to instruct the corner of the document P to abut. In this embodiment, the corner of the document table 11a to which the mark MK is attached is defined as “lower left corner”, and the corner of the document table 11a located at a point farther in the main scanning direction than the lower left corner is defined. The “lower right corner” is defined, and the corner of the document table 11a located at a point away from the lower left corner in the sub-scanning direction is defined as “upper left corner”.

  That is, in the multifunction machine 1, the lower left corner of the document table 11a (in other words, the lower left corner inside the frame 13a) is determined as the position where the corner of the document P should be aligned, and the left end extending from the lower left corner of the document table 11a. The edge and the lower edge are defined as edges to which the side (edge) of the document P should be aligned, and the multi-function device 1 indicates that the user generally places the document P in alignment with the mark MK. Under the assumption, the manuscript size is estimated, and the manuscript copy operation is realized.

  In addition, in the multifunction machine 1 of this embodiment, the lower left corner of the readable area R0 corresponding to the inner lower left corner of the frame 13a with the mark MK is the origin, the main scanning direction is the X axis, and the sub scanning direction is An XY coordinate system (see FIG. 2B) with the Y axis is introduced, and the CPU 70 executes the copy control process shown in FIG. 3 using this XY coordinate system.

  Next, a copy control process executed by the CPU 70 will be described. FIG. 3 is a flowchart showing a copy control process executed by the CPU 70 when a copy command is input through the operation keys.

  When the copy control process is started, the CPU 70 determines whether or not the copy command input through the operation key is an “automatic scaling / tilt correction” copy command (S110), and copies the “automatic scaling / tilt correction”. If it is determined that it is a command (Yes in S110), the automatic zoom copying process shown in FIG. 4 is executed in S120. Thereafter, the copy control process ends.

  On the other hand, if it is determined that the copy command input through the operation key is a copy command other than the “automatic scaling / tilt correction” copy command (for example, a normal copy command without an automatic scaling or tilt correction instruction). In step S130, the CPU 70 executes processing corresponding to the input copy command, and then ends the copy control processing.

Next, the automatic scaling copying process executed by the CPU 70 in S120 will be described. FIG. 4 is a flowchart showing the automatic scaling copying process executed by the CPU 70.
When the CPU 70 starts the automatic scaling copying process in S120, first, the document reading start position is set to the lower end (Y = 0) of the readable area R0, and the document reading end position is predetermined in the design stage. By setting a position (Y = YPRE) that is a predetermined distance away from the lower end of the readable area R0, a part of the readable area R0 corresponding to the area from the original reading start position to the original reading end position is read. An area (in other words, a pre-scan area) is set, and the image reading unit 10 is caused to execute a pre-scan operation on the reading area through the reading control unit 20 (S210: tip portion pre-scanning process).

  That is, in S210, the pre-scanning operation is not performed on the entire readable area R0 by the image reading unit 10, but from the lower end of the readable area R0 to a point a predetermined distance away from the readable area R0 (this embodiment). In the example, the image reading unit 10 is caused to perform a pre-scan operation for a point 30 mm away from the lower end of the readable region R0. As a result, the CPU 70 acquires, as a pre-scan result, image data representing the read result of a partial area of the readable area R0 from the image reading unit 10.

  As a pre-scan operation, the image reading unit 10 conveys the reading unit 15 in the sub-scanning direction at a speed corresponding to the pre-scan resolution from the end of the set reading area to the end of the reading area. The image is read by the reading unit 15 and, as a prescan result, low resolution image data (hereinafter referred to as “prescan image data”) is recorded in the RAM 80 through the read control unit 20.

  When the processing in S210 is completed in this way, the CPU 70 next performs edge detection processing on the pre-scan image data recorded in the RAM 80, and generates edge image data corresponding to this image data ( S215). That is, the prescan image data recorded in the RAM 80 is passed through an image filter for edge detection (a well-known differential filter), and edge image data representing an edge image corresponding to the image data is generated.

  When the process in S215 is completed, the CPU 70 sets the generated edge image data as inspection target data, and executes the document estimation process shown in FIG. 5 (S220). Although details will be described later, in this document estimation process, the inspection target data (edge image data obtained in S215) is analyzed to detect the document edge, and the document placed on the document table 11a is detected from the detection result. Size (vertical width and horizontal width), the inclination angle θ of the original, and the original placement area on the original table 11a.

  When the processing in S220 is completed, the CPU 70 causes the printing unit 30 to perform a paper feeding operation through the print control unit 40 (S230), and detects the size of the recording paper that is the target of the paper feeding operation (S230). S240). Thereafter, the process proceeds to S251. The size of the recording paper can be detected by a well-known method using, for example, a sensor provided in the recording paper conveyance path.

  In step S251, the CPU 70 sets a scaling ratio according to a predetermined calculation formula based on the document size estimated in step S220 and the recording paper size detected in step S240. Specifically, the scaling factor is set to a magnification corresponding to the ratio between the document size and the recording paper size (for example, recording paper short side length ÷ original short side length), and the copy of the original is performed in the subsequent processing. The image is enlarged (or reduced if the magnification is less than 1) at a magnification corresponding to the ratio of the recording paper size to the document size and printed on the recording paper.

  When this processing is completed, the CPU 70 sets the inclination correction amount of the image data from the original inclination angle θ estimated in S220 (S253). Specifically, the tilt correction amount is set so that the copy image of the document is printed straight on the recording paper without tilting. However, when the estimated document inclination angle θ is a minute amount (in the present embodiment, −0.5 degrees ≦ θ ≦ 0.5 degrees), the inclination correction amount is set to zero in consideration of errors. To do.

  When the processing in S253 is completed, the CPU 70 sets the document reading start position and the document reading end position based on the document placement area information estimated in S220, and sets the document reading start position and the document reading end position. The determined reading area is made to correspond to the document placement area on the document table 11a (S255).

  Specifically, the document reading start position is set to a position corresponding to the end point of the document located on the lowermost side in the sub-scanning direction so that the reading unit 15 can read the entire document placement area. The document reading end position is set to a position corresponding to the end point of the document located on the uppermost side in the sub-scanning direction, and the reading area is made to correspond to the document placement area on the document table 11a. When this process is finished, the process proceeds to S270.

  In step S <b> 270, the CPU 70 controls the image reading unit 10 through the reading control unit 20 to convey the reading unit 15 to the image reading unit 10 from the document reading start position to the document reading end position in the sub-scanning direction. In addition, during the conveyance, the reading unit 15 performs a reading operation for each line, causes the reading unit 15 to read the image of the set reading area, and image data representing a reading result of the reading area is It is recorded in the RAM 80 (S270).

  When this process is completed, the CPU 70 enlarges or reduces the image data representing the read result recorded in the RAM 80 at a preset magnification, and performs a rotation process for a preset inclination correction amount. Then, the image data representing the read result is converted into image data for printing, and the converted image data is set as print target data (S280). However, if the tilt correction amount is set to zero, the rotation process is not executed in S280.

  When the process in S280 is completed, the CPU 70 executes a print process for the print target data (S290). That is, the printing unit 30 causes the printing unit 30 to print an image based on the print target data on the fed recording paper. Thereafter, the copy control process ends.

  In this way, in the automatic scaling copying process, the document edge is detected from the prescan result of a partial area in the readable area R0, the document size and the like are estimated, and based on the result, during the main scan (S270 execution) ) And a variable magnification and an inclination correction amount are determined. Then, the copy image of the original is enlarged or reduced to a size corresponding to the ratio between the original size and the recording paper size, and the inclination is corrected and printed on the recording paper.

Subsequently, the document estimation process executed by the CPU 70 in S220 will be described with reference to FIG. FIG. 5 is a flowchart showing document estimation processing executed by the CPU 70.
When the document estimation process is started, the CPU 70 first sets the variable S to YPRE which is the Y coordinate of the upper end of the pre-scan area (S = YPRE), thereby setting the inspection range in the Y-axis direction to Y = 0. To Y = S = YPRE (S410). When this process ends, the process proceeds to S420, and a right edge detection process is executed. FIG. 6 is a flowchart showing the right edge detection process executed by the CPU 70.

  When the right edge detection process is started, the CPU 70 first sets the sub-scanning direction inspection position Ye to the lower end Y coordinate of the readable area R0 (Ye = 0) in S510, and the main scanning direction inspection position in S515. Xe is set to the right end X coordinate of the readable region R0 (Xe = XMAX). In S515, the variable C is initialized to zero (C = 0). Thereafter, it is determined whether or not the sub-scanning direction inspection position Ye exceeds the inspection range (S520). Specifically, it is determined whether or not Ye> S.

  When determining that the sub-scanning direction inspection position Ye does not exceed the inspection range (No in S520), the CPU 70 determines the inspection position (X, Y) = determined by the main scanning direction inspection position Xe and the sub-scanning direction inspection position Ye = It is determined whether or not the inspection position (Xe, Ye) is an edge point by referring to the pixel value of the inspection target data corresponding to (Xe, Ye) (S525). FIG. 7 is an explanatory diagram showing the configuration of the edge image data, and further showing the locus of the edge point tracked by the right edge detection process. In the example illustrated in FIG. 7, a point with a pixel value “1” corresponds to an edge point.

  If the CPU 70 determines that the inspection position (Xe, Ye) is not an edge point (No in S525), the CPU 70 proceeds to S540 and moves the main scanning direction inspection position Xe by one pixel in the X-axis minus direction. (Xe ← Xe−1), and it is determined whether or not the updated main scanning direction inspection position Xe protrudes from the left end of the readable area R0. Specifically, it is determined whether or not Xe <0 (S543).

  If the CPU 70 determines that the main scanning direction inspection position Xe does not protrude from the left end of the readable region R0 (No in S543), the CPU 70 proceeds to S520, and the sub scanning direction inspection position Ye exceeds the inspection range. If it is determined whether or not (Ye> S) and the sub-scanning direction inspection position Ye does not exceed the inspection range (No in S520), the process proceeds to S525 and the sub-scanning direction inspection position Ye is determined. If it is determined that the value exceeds the inspection range (Yes in S520), the process proceeds to S590.

  On the other hand, in S543, when it is determined that the main scanning direction inspection position Xe protrudes from the left end of the readable area R0 (when it is determined that Xe <0), the CPU 70 proceeds to S547, and the main scanning direction inspection position Xe. Is set to the right end (XMAX) of the readable region R0 (Xe = XMAX), and the sub-scanning direction inspection position Ye is updated to a value obtained by adding 8 to the current value Ye (Ye ← Ye + 8). That is, the sub-scanning direction inspection position Ye is set to a position advanced by 8 pixels in the Y-axis direction. Thereafter, the process proceeds to S520 and the above-described processing is executed.

  If the CPU 70 determines that the inspection position (Xe, Ye) is an edge point (Yes in S525), the CPU 70 proceeds to S530, and the coordinates of the current inspection position (Xe, Ye) are set in the variable (X0, Y0). The value is set (X0 ← Xe, Y0 ← Ye). Further, the variable Y1 is updated to a coordinate value Y1 (= Y0 + 8) advanced by 8 pixels from the coordinate Y = Y0 in the Y-axis direction (S533).

  Thereafter, it is determined whether at least one of the point of coordinates (X0-1, Y1), the point of coordinates (X0, Y1), and the point of coordinates (X0 + 1, Y1) is an edge point (S537). If none of the point of (X0-1, Y1), the point of coordinates (X0, Y1), and the point of coordinates (X0 + 1, Y1) is an edge point (No in S537), the process proceeds to S540. The scanning direction inspection position Xe is updated to a position moved by one pixel in the X axis minus direction (Xe ← Xe−1).

  On the other hand, when it is determined that at least one of the point of coordinates (X0-1, Y1), the point of coordinates (X0, Y1), and the point of coordinates (X0 + 1, Y1) is an edge point (Yes in S537), The process proceeds to S550, where it is determined that the point of coordinates (X0, Y0) is a continuous edge point, and this coordinate (X0, Y0) is temporarily stored as coordinate data of the continuous edge point. .

  When this processing is finished, the CPU 70 proceeds to S560, and at the point of the coordinates (X0-1, Y1), the coordinates (X0, Y1), and the coordinates (X0 + 1, Y1) at the edge point. One of the points is selected according to a predetermined priority.

  Specifically, in the right edge detection process, higher priority is set for coordinates closer to the right end side of the readable area R0. That is, the point of the coordinate (X0 + 1, Y1) close to the right end side of the readable area R0 is set to the priority “high”, the point of the coordinate (X0, Y1) is set to the priority “medium”, and the coordinate (X0− 1, Y1) is set to the priority “small”.

  The priority is set in this way because nothing exists on the platen 11a on the right side from the right end of the document, and there should be no edge point, and an edge point close to the right end side of the readable area R0. This is because there is a higher possibility that the edge point corresponds to the right edge of the document.

  That is, in S560, according to the above priority, the point of the coordinate (X0-1, Y1), the point of the coordinate (X0, Y1), and the point of the coordinate (X0 + 1, Y1) are points that are edge points. To select the point with the highest priority. Then, the variable X0 is updated to the X coordinate of the selected point, and the variable Y0 is updated to the Y coordinate of the selected point. Thereafter, the variable C is updated to a value obtained by adding 1 (S563).

  When this process is finished, the CPU 70 determines whether or not the value of the updated variable C is C = 8. If the CPU 70 determines that C = 8 is not satisfied (No in S567), the process proceeds to S533. The variable Y1 is updated to the Y coordinate of the point advanced by 8 pixels in the Y-axis direction from the coordinate Y = Y0 (Y1 = Y0 + 8). Thereafter, the processing after S537 is executed.

  By executing such processing, the CPU 70 checks whether the edge points are continuous in the Y-axis direction every 8 pixels in the Y-axis direction, as shown in FIG. If it is determined that C = 8 (Yes in S567), the coordinate data of a total of eight points determined as the continuous edge points in the process of S550 that has been repeated eight times is stored in the RAM 80 as the original right edge data. (S570).

When this process is finished, the CPU 70 updates the sub-scanning direction inspection position Ye to a position advanced by 64 pixels (S580), and proceeds to S515.
In this way, the CPU 70 repeatedly detects eight consecutive edge points as one set. When the sub-scanning direction detection position Ye exceeds the inspection range (Yes in S520), the process proceeds to S590, and the right edge (represents the right edge of the document) is selected from the coordinate data group stored as document right edge data in S570. The coordinate data group with low accuracy is deleted as the edge point), and the document right edge data is determined.

  Specifically, in S590, the original right edge data is set as processing target data, and the first edge determination process shown in FIG. 8 is executed, whereby a coordinate data group with low accuracy is selected as the right edge from the original right edge data. Delete and confirm the document right edge data. FIG. 8 is a flowchart showing the first edge determination process executed by the CPU 70.

  When the first edge determination process is started, the CPU 70 determines whether or not the preset processing target data includes coordinate data of at least eight edge points (S610), and the eight processing target data are included. If it is determined that the coordinate data of the edge point is included (Yes in S610), the process proceeds to S620, and the processing target data does not include the coordinate data of eight or more edge points (in other words, the processing target data is not included). If it is determined that the data is empty data (No in S610), the first edge determination process is terminated without executing the processes of S620 to S670.

  On the other hand, when proceeding to S620, the CPU 70 sets a block having the maximum Y coordinate value as the inspection block in the processing target data based on the coordinate data group of the edge points constituting the processing target data.

  However, here, the edge detection processing (right edge detection processing or left edge detection processing or lower edge detection) that is a set of edge points in which coordinate data is registered in the processing target data and has led to generation of the processing target data. A set of eight edge points determined to be continuous in (processing) is referred to as a “block”.

  As described above, in this embodiment, in the right edge detection process, eight edge points continuous in the Y-axis direction are detected as one set, and the coordinate data of the detected eight edge points is recorded in the RAM 80. (See S570) The next set of consecutive edge points is searched again from the right end of the image data to the inside of the image data (see S580 and S515).

  In S620, a block consisting of a set of eight consecutive edge points whose coordinate data has been registered in the RAM 80 in this way and including an edge point having the maximum Y coordinate value as an element is extracted from the processing target data. Select it and set it to the test block.

  In addition, when the processing in S620 is completed in this way, the CPU 70 proceeds to S630 and determines whether or not the coordinate data of the edge point adjacent in the Y axis minus direction from the inspection block is included in the processing target data. to decide.

  Here, if the pixel position of the edge point having the smallest Y coordinate value among the edge points constituting the inspection block is (X, Y) = (X0, Y0), the edge point adjacent in the Y axis minus direction It is determined whether or not the coordinate data of the edge point having the Y coordinate value of Y = Y0-8 is included in the processing target data. Thereby, it is determined whether or not there is an adjacent edge point (in other words, an adjacent block) on the one-dimensional space only on the Y axis.

  If it is determined that the coordinate data of the edge point adjacent in the Y axis minus direction is included in the processing target data (Yes in S630), the process proceeds to S640, and the coordinate data of the edge point adjacent in the Y axis minus direction is obtained. If it is determined that the data is not included in the processing target data (No in S630), the process proceeds to S670.

  In step S640, the CPU 70 determines the X coordinate value of the edge point at the end of the inspection block in the Y-axis minus direction (in other words, the X coordinate of the edge point having the smallest Y coordinate value among the edge points constituting the inspection block). Value) and the X coordinate value of the edge point adjacent in the negative Y-axis direction from the above-described inspection block. The pixel position of the edge point at the end of the inspection block in the Y-axis minus direction is (X, Y) = (X0, Y0), and the pixel position of the edge point adjacent to the inspection block in the Y-axis minus direction is (X, Y). = (X1, Y0-8), δX = | X0−X1 | is calculated.

  When this process is completed, the CPU 70 determines whether or not the difference δX is less than a threshold value determined in advance in the design stage (S650). If the CPU 70 determines that the difference δX is less than the threshold value (Yes in S650), the inspection is performed. The block is updated to the next block adjacent in the Y-axis minus direction (S660), and then the process proceeds to S630. For example, when the pixel position of the edge point at the end of the Y-axis minus direction of the inspection block is (X, Y) = (X0, Y0), the Y coordinate value ranges from Y = Y0-8 to Y = Y0-64. After a set of a total of eight edge points that take values in the range is set as a new inspection block, the process proceeds to S630.

  On the other hand, when determining that the difference δX is equal to or larger than the threshold (No in S650), the CPU 70 determines the accuracy of the edge point located in the Y-axis minus direction from the end of the current inspection block in the Y-axis minus direction as the document edge. The process moves to S670, and all the coordinate data of the edge points located in the Y-axis minus direction from the end of the Y-axis minus direction of the current inspection block are deleted from the process target data, and the process target data is Update.

  For example, if the pixel position of the edge point at the end in the Y-axis minus direction of the current inspection block is (X, Y) = (X0, Y0), all the coordinate data of the edge point whose Y coordinate value is less than Y0 Delete the processing target data and update the processing target data. Thereafter, the first edge determination process ends. In the first edge determination process, the coordinate data having low accuracy as the document edge is deleted from the coordinate data of the edge points constituting the processing target data in this way.

  That is, in S590 of the right edge detection process, the coordinate data of the edge point with low accuracy is deleted as the right edge from the document right edge data in this way, and the document right edge data is determined.

  FIG. 9 is an explanatory diagram showing a method for determining the right edge data of the document in S590. FIG. 9B shows calculation of δX together with edge image data that represents every eight pixels in the Y-axis direction. It is explanatory drawing which represented the method visually.

  In S590, as described above, an edge point having a larger Y coordinate value is treated as having a higher probability that the edge point is an edge point (right edge) representing the right edge of the document. The reason why the edge point group having a small Y coordinate value existing at a position greatly displaced in the axial direction is deleted from the original right edge data is that the inspection positions (Xe, Ye) are sequentially detected from Ye = 0 in the right edge detection process. When the edge point representing the right edge of the document is searched by shifting, if the document is tilted to the left, it corresponds not to the right edge of the document but to the ruled line drawn inside the document as shown in FIG. This is because the edge point to be detected may be erroneously detected as the edge point representing the right edge of the document.

  Such an erroneous detection is likely to occur at an edge point group with a small Y coordinate value located near the lower end of the document table 11a, and an edge point with a large Y coordinate value located near the center of the document table 11a. It is unlikely that it is occurring in a group. For this reason, in S590, an edge point with a larger Y coordinate value is treated as having a higher probability that the edge point is an edge point (right edge) representing the right edge of the document, and an edge point with a lower accuracy is treated as the right edge of the document. The data is deleted from the data, and the right edge data of the document is determined.

  In addition, when the processing in S590 is completed in this way, the CPU 70 ends the right edge detection processing in S420 and proceeds to S430, and in S430, the detection of the right edge fails in the immediately preceding right edge detection processing. Determine whether or not. Specifically, when there is no coordinate data registered as document right edge data in the immediately preceding right edge detection process, it is determined that detection of the right edge has failed, and there is coordinate data registered as document right edge data It is determined that the right edge has been successfully detected.

  However, exceptionally, when the coordinate data registered as the document right edge data is in the vicinity of the left end (Y = 0) of the readable region R0, the coordinate data of the left edge (the edge point representing the left end of the document) is It may be determined that the detection of the right edge has failed even when there is coordinate data registered as the right edge data of the document, assuming that it is erroneously registered as the right edge data of the document.

  Examples of cases where the detection of the right edge fails include a case where the document size is larger than the size of the document table 11a and the right edge of the document protrudes from the document table 11a.

  If it is determined that the detection of the right edge has failed (Yes in S430), the CPU 70 proceeds to S431 and executes error processing. For example, the multi-function device 1 displays, as an error process in S431, a message prompting the user to input another copy command because the document is placed correctly or the automatic scaling copying function cannot be used. After the display, the process can be configured to execute the process of forcibly terminating the copy control process corresponding to the copy command that led to the execution of S431 and invalidating the copy command.

  In addition, the multi function device 1 performs the error processing in S431 as follows: “The size of the document placed on the document table 11a is formally estimated to be a predetermined fixed size, and the inclination angle θ of the document is determined. Further, it is estimated that the original is placed on the area where the original is placed when it is assumed that the lower left corner of the estimated size of the original is correctly abutted against the lower left corner of the original table 11a. The “estimating” process may be executed, and then the document estimation process may be terminated and the process may proceed to S230.

  On the other hand, if it is determined in S430 that the right edge has been successfully detected (No in S430), the CPU 70 proceeds to S433 and linearly approximates each point indicated by the document right edge data determined in the right edge detection process. An approximate straight line of the point group indicated by the original right edge data (hereinafter referred to as “right edge approximate straight line”) is calculated.

  When this process is finished, the CPU 70 proceeds to S440 and determines whether or not the right edge approximate straight line is inclined more than a predetermined angle with respect to the Y axis. In this embodiment, specifically, it is determined whether or not the right edge approximate straight line is inclined more than 0.5 degrees with respect to the Y axis.

  If it is determined that the slope of the right edge approximate straight line is equal to or smaller than the predetermined angle (No in S440), the CPU 70 is placed on the document table 11a by being abutted against the inner left lower corner of the frame 13a. As shown in the right diagram of FIG. 5, the intersection of the right edge approximate straight line and the lower edge of the document table 11a is estimated as the lower right corner position of the document (S441).

  When this processing is completed, the CPU 70 estimates that the lower left corner position of the document is the lower left corner point of the document table 11a (S443), and based on the position coordinates of the angular position estimated in S441 and S443, the document table. Under the assumption that the document placed on the document 11a is a square-shaped standard paper, the size of the document placed on the document table 11a and the placement area of the document on the document table 11a are estimated, and the document Is estimated (S445).

  Specifically, the length from the lower left corner position to the lower right corner position of the document estimated in S441 and S443 is estimated as the short side length (horizontal width) of the document, and √2 times the short side length is The document size is estimated by estimating the long side length (vertical width) of the document, and the document of the document size is placed on the document table 11a starting from the estimated lower left corner position and lower right corner position. As an example, the coordinates at which the upper left corner and the upper right corner of the document are located are derived, and an area within a quadrilateral formed by connecting these four corners is estimated as the document placement area. Here, the reason why √2 times the original short side length is estimated as the long side length of the original is that the ratio of the short side length to the long side length of the standard paper is 1: √2. This is because.

  Further, in S441 and S443, the angular position is obtained on the assumption that the document is accurately abutted against the lower left corner of the document table 11a and placed on the document table 11a. Therefore, in S445, the inclination angle of the document is determined. It is estimated that θ is zero (no slope).

  When the process in S445 is completed in this way, the CPU 70 ends the document estimation process (S220), proceeds to S230, and executes the subsequent process as described above. That is, based on the document size estimated by the document estimation process, the document tilt angle, and the document placement area information, the scaling factor and the tilt correction amount are set (S251, S253), and the reading area is further set (S255). ). Thereafter, the process shifts to the main scan (S270), and the copy operation is realized. However, since the inclination angle θ of the document is estimated to be zero here, the inclination correction is not actually executed.

  When the story is returned to S440 (see FIG. 5) of the document estimation process, the CPU 70 determines that the right edge approximate straight line is inclined more than a predetermined angle with respect to the Y axis (Yes in S440), the process proceeds to S450. Then, the left edge detection process shown in FIG. 10 is executed. FIG. 10 is a flowchart showing the left edge detection process executed by the CPU 70.

  When the left edge detection process is started, the CPU 70 sets the sub-scanning direction inspection position Ye = 0 in S710, and sets the main scanning direction inspection position Xe to the left end X coordinate of the readable area R0 in S715. (Xe = 0). In S715, the variable C is initialized to zero (C = 0).

  Thereafter, the CPU 70 determines whether or not the sub-scanning direction inspection position Ye exceeds the inspection range (that is, whether or not Ye> S) (S720), and the sub-scanning direction inspection position Ye exceeds the inspection range. If it is determined that it is not (No in S720), the inspection position (Xe, Ye) is an edge point with reference to the pixel value of the inspection target data corresponding to the inspection position (X, Y) = (Xe, Ye). Whether or not (S725).

  If it is determined that the inspection position (Xe, Ye) is not an edge point (No in S725), the process proceeds to S740, and the main scanning direction inspection position Xe is updated to a point moved by one pixel in the X axis plus direction ( Xe ← Xe + 1), it is determined whether or not the updated main scanning direction inspection position Xe protrudes from the right end of the readable area R0 (S743). Specifically, it is determined whether Xe> XMAX (S743). If it is determined that the main scanning direction inspection position Xe does not protrude from the right end of the readable area R0 (No in S743), the process proceeds to S720.

  On the other hand, when it is determined in S743 that the main scanning direction inspection position Xe protrudes from the right end of the readable region R0 (when it is determined that Xe> XMAX), the CPU 70 proceeds to S747 and sets the main scanning direction inspection position Xe. The sub scanning direction inspection position Ye is set to a position advanced by 8 pixels in the Y-axis direction (Y ← Ye + 8) while being set to the left end of the readable region R0 (Xe = 0). Thereafter, the process proceeds to S720.

  If the CPU 70 determines that the inspection position (Xe, Ye) is an edge point (Yes in S725), the CPU 70 proceeds to S730 and sets the variable (X0, Y0) to the current inspection position (Xe, Ye). The coordinate value is set (X0 ← Xe, Y0 ← Ye), and the variable Y1 is updated to the value Y1 = Y0 + 8 (S733).

  Thereafter, it is determined whether at least one of the point of coordinates (X0-1, Y1), the point of coordinates (X0, Y1), and the point of coordinates (X0 + 1, Y1) is an edge point (S737). If none of the point of (X0-1, Y1), the point of coordinates (X0, Y1), and the point of coordinates (X0 + 1, Y1) is an edge point (No in S737), the process proceeds to S740.

  On the other hand, when it is determined that at least one of the point of coordinates (X0-1, Y1), the point of coordinates (X0, Y1), and the point of coordinates (X0 + 1, Y1) is an edge point (Yes in S737), Moving to S750, it is determined that the point of coordinates (X0, Y0) is a continuous edge point, and the coordinates (X0, Y0) are temporarily stored as coordinate data of the continuous edge point. .

  When this process is finished, the CPU 70 proceeds to S760, where the coordinates (X0-1, Y1), the coordinates (X0, Y1), and the coordinates (X0 + 1, Y1) are the edge points. One of the points is selected according to a predetermined priority.

  Specifically, in the left edge detection process, higher priority is set for coordinates closer to the left end side of the readable area R0. That is, the point of the coordinates (X0-1, Y1) close to the left end side of the readable area R0 is set to the priority “large”, the point of the coordinates (X0, Y1) is set to the priority “medium”, and the coordinates ( The point of X0 + 1, Y1) is set to the priority “small”. The priority is set in this way because there is nothing on the document table 11a on the left side of the left edge of the document, and there should be no edge point, and the edge point close to the left edge side of the readable area R0. This is because there is a high possibility that the edge point is an edge point corresponding to the left end of the document.

  Therefore, in S760, according to the priority, the point of the coordinate (X0-1, Y1), the point of the coordinate (X0, Y1), and the point of the coordinate (X0 + 1, Y1) are points that are edge points. To select the point with the highest priority. Then, the variable X0 is updated to the X coordinate of the selected point, and the variable Y0 is updated to the Y coordinate of the selected point. Thereafter, the variable C is updated to a value obtained by adding 1 (S763).

  When this process is finished, the CPU 70 determines whether or not the value of the updated variable C is C = 8. If the CPU 70 determines that C = 8 is not satisfied (No in S767), the process proceeds to S733. The variable Y1 is updated to Y1 = Y0 + 8. Thereafter, the processing after S737 is executed.

  By such a processing procedure, the CPU 70 checks whether the edge points are continuous in the Y-axis direction every 8 pixels in the Y-axis direction, as in the right edge detection process (see FIG. 7). If it is determined that C = 8 (Yes in S767), the coordinate data of a total of eight points determined as continuous edge points in S750 are stored in the RAM 80 as document left edge data (S770).

When this process is completed, the CPU 70 proceeds to S780, updates the sub-scanning direction inspection position Ye to a point advanced by 64 pixels, and then proceeds to S715.
In this way, the CPU 70 repeatedly detects eight consecutive edge points. When the sub-scanning direction detection position Ye exceeds the inspection range (Yes in S720), the process proceeds to S790, and the left edge (representing the left end of the document) is selected from the coordinate data group stored as document left edge data in S770. The coordinate data group with low accuracy is deleted as the edge point), and the document left edge data is determined.

  Specifically, similar to the processing in S590, the left edge data of the document is set as processing target data, and the first edge determination processing shown in FIG. Delete from the document left edge data and determine the document left edge data.

  In addition, when the left edge detection process is completed in this way, the CPU 70 proceeds to S455, linearly approximates each point indicated by the document left edge data determined by the left edge detection process, and the document left edge data is obtained. An approximate straight line (hereinafter referred to as “left edge approximate straight line”) of the point group shown is calculated. Thereafter, the process proceeds to S460.

  However, if there is no coordinate data registered as document left edge data in the immediately preceding left edge detection process, it is determined that detection of the left edge has failed, and the process proceeds to S460 without calculating the left edge approximate straight line. It shall be. In addition, when all or part of the coordinate data registered as the document left edge data is also registered in the document right edge data, the coordinate data of the right edge (the edge point representing the right edge of the document) is incorrect. Even if there is coordinate data registered as document left edge data, it is determined that detection of the left edge has failed, and without calculating the left edge approximate straight line, You may make it transfer to S460.

In S460, the CPU 70 executes the lower edge detection process shown in FIG. FIG. 11 is a flowchart showing the lower edge detection process executed by the CPU 70.
When the lower edge detection process is started, the CPU 70 sets the main scanning direction inspection position Xe to the left end X coordinate of the readable area R0 (Xe = 0) in S810, and the sub scanning direction inspection position Ye in S815. Is set to the lower end Y coordinate of the readable area R0 (Ye = 0). In S815, the variable C is initialized to zero (C = 0).

  When this process is finished, the CPU 70 determines whether or not the main scanning direction inspection position Xe exceeds the right end of the readable area R0 (that is, whether Xe> XMAX) (S820). If it is determined that the scanning direction inspection position Xe does not exceed the right end of the readable region R0 (No in S820), the pixel value of the inspection target data corresponding to the inspection position (X, Y) = (Xe, Ye) is referred to. Then, it is determined whether or not the inspection position (Xe, Ye) is an edge point (S825).

  If it is determined that the inspection position (Xe, Ye) is not an edge point (No in S825), the process proceeds to S840, and the sub-scanning inspection position Ye is updated to a position moved by one pixel in the Y-axis plus direction. (Ye ← Ye + 1), it is determined whether or not the updated sub-scanning direction inspection position Ye exceeds the inspection range (whether or not Ye> S) (S843).

If it is determined that the sub-scanning direction inspection position Ye does not exceed the inspection range (No in S843), the process proceeds to S820.
On the other hand, when it is determined in S843 that the sub-scanning direction inspection position Ye exceeds the inspection range (when it is determined that Ye> S), the CPU 70 proceeds to S847, and the sub-scanning direction inspection position Ye is set in the readable area. While setting at the lower end of R0 (Ye = 0), the main scanning direction inspection position Xe is set to a position advanced by 8 pixels in the X-axis direction (Xe ← Xe + 8). Thereafter, the process proceeds to S820.

  In addition, when the CPU 70 determines in S825 that the inspection position (Xe, Ye) is an edge point (Yes in S825), the CPU 70 proceeds to S830 and sets the current inspection position (Xe, Y0) in the variable (X0, Y0). The coordinate value of (Ye) is set (X0 ← Xe, Y0 ← Ye). Also, the variable X1 is updated to the value X1 = X0 + 8 (S833).

  Thereafter, the CPU 70 determines whether or not at least one of the coordinates (X1, Y0-1), the coordinates (X1, Y0), and the coordinates (X1, Y0 + 1) is an edge point (S837). ), If any of the point of coordinates (X1, Y0-1), the point of coordinates (X1, Y0), and the point of coordinates (X1, Y0 + 1) is not an edge point (No in S837), the process proceeds to S840. .

  On the other hand, if it is determined that at least one of the coordinates (X1, Y0-1), the coordinates (X1, Y0), and the coordinates (X1, Y0 + 1) is an edge point (Yes in S837), the CPU 70 Moving to S850, it is determined that the point of coordinates (X0, Y0) is a continuous edge point, and this coordinate (X0, Y0) is temporarily stored as coordinate data of a continuous edge point. .

  When this process is finished, the CPU 70 proceeds to S860, where the coordinates (X1, Y0-1), the coordinates (X1, Y0), and the coordinates (X1, Y0 + 1) are the edge points. One of the points is selected according to a predetermined priority.

  Specifically, in the lower edge detection process, a higher priority is set for coordinates closer to the lower end side of the readable area R0. That is, the point of the coordinates (X1, Y0-1) close to the lower end side of the readable area R0 is set to the priority “large”, the point of the coordinates (X1, Y0) is set to the priority “medium”, and the coordinates ( The point of (X1, Y0 + 1) is set to the priority “small”. The priority is set in this way because there is nothing on the document table 11a below the lower end of the original, there should be no edge point, and the edge close to the lower end of the readable area R0. This is because the point is more likely to be the edge point corresponding to the lower end of the document.

  Therefore, in S860, according to this priority, the point that is the edge point among the point of the coordinate (X1, Y0-1), the point of the coordinate (X1, Y0), and the point of the coordinate (X1, Y0 + 1) is preferred. Select the point with the highest degree. Then, the variable X0 is updated to the X coordinate of the selected point, and the variable Y0 is updated to the Y coordinate of the selected point. Thereafter, the variable C is updated to a value obtained by adding 1 (S863).

  When this process is finished, the CPU 70 determines whether or not the value of the updated variable C is C = 8 (S867). If it is determined that C = 8 is not satisfied (No in S867), the process proceeds to S833. Then, the variable X1 is updated to X1 = X0 + 8. Thereafter, the processing after S837 is executed.

  By executing such processing, the CPU 70 checks whether the edge points are continuous in the X-axis direction every 8 pixels in the X-axis direction. If it is determined that C = 8 (Yes in S867), the coordinate data of a total of eight points determined as continuous edge points in S850 are stored in the RAM 80 as document lower edge data (S870).

When this process is finished, the CPU 70 proceeds to S880, updates the main scanning direction inspection position Xe to a point advanced by 64 pixels, and then proceeds to S815.
In this way, the CPU 70 repeatedly detects eight consecutive edge points. When the main scanning direction detection position Xe exceeds the right end of the readable area R0 (Yes in S820), the process proceeds to S890, and the lower edge (original of the document) is selected from the coordinate data group stored as the document lower edge data in S870. The coordinate data group with low accuracy is deleted as the edge point representing the lower end), and the document lower edge data is determined.

  Specifically, in S890, the lower edge data of the document is set as processing target data, and the second edge determination process is executed to delete the coordinate data group having low accuracy as the lower edge from the lower edge data of the document. Confirm the bottom edge data. FIG. 12 is a flowchart showing the second edge determination process executed by the CPU 70.

  When the second edge determination process is started, the CPU 70 determines whether or not the preset processing target data includes coordinate data of at least eight edge points (S910), and the eight processing target data are included. If it is determined that the coordinate data of the edge point is included (Yes in S910), the process proceeds to S920, and the processing target data does not include coordinate data of eight or more edge points (in other words, the processing target data is not included). If it is determined that the data is empty data (No in S910), the second edge determination process is terminated without executing the processes of S920 to S970.

  On the other hand, when proceeding to S920, the CPU 70 sets the block having the maximum X coordinate value in the processing target data as the inspection block based on the coordinate data group of the edge points constituting the processing target data.

  As described above, in this embodiment, in the lower edge detection process, eight edge points continuous in the X-axis direction are detected as one set, and the coordinate data of the detected eight edge points is recorded in the RAM 80 ( In step S870, the next set of consecutive edge points is searched again from the lower end of the image data (inspection data) toward the inside of the image data (see steps S880 and S815).

  In S920, a block composed of a set of eight consecutive edge points whose coordinate data is registered in the RAM 80 in this way and including an edge point having the maximum X coordinate value as an element is extracted from the processing target data. Select it and set it to the test block.

  In addition, when the processing in S920 is completed in this way, the CPU 70 proceeds to S930 and determines whether or not the coordinate data of the edge point adjacent in the negative direction of the X axis from the inspection block is included in the processing target data. to decide.

  Here, if the pixel position of the edge point having the smallest X coordinate value among the edge points constituting the inspection block is (X, Y) = (X0, Y0), the edge point adjacent in the X axis minus direction It is determined whether or not the coordinate data of the edge point whose X coordinate value is X = X0-8 is included in the processing target data. Thereby, the presence / absence of an adjacent edge point (in other words, an adjacent block) is determined on the one-dimensional space only on the X axis.

  When it is determined that the coordinate data of the edge point adjacent in the X axis minus direction is included in the processing target data (Yes in S930), the process proceeds to S940, and the coordinate data of the edge point adjacent in the X axis minus direction is obtained. If it is determined that the data is not included in the processing target data (No in S930), the process proceeds to S970.

  In S940, the CPU 70 determines the Y coordinate value of the edge point at the end of the X-axis minus direction of the inspection block (in other words, the Y coordinate of the edge point having the smallest X coordinate value among the edge points constituting the inspection block). Value) and the Y coordinate value of the edge point adjacent in the X-axis minus direction from the above-described inspection block. The pixel position of the edge point at the end of the X axis minus direction of the inspection block is (X, Y) = (X0, Y0), and the pixel position of the edge point adjacent in the X axis minus direction from the inspection block is (X, Y). = (X0-8, Y1), δY = | Y0-Y1 | is calculated.

  When this process is completed, the CPU 70 determines whether or not the difference δY is less than a threshold value determined in advance in the design stage (S950). If the CPU 70 determines that the difference δY is less than the threshold value (Yes in S950), the inspection is performed. The block is updated to the next block adjacent in the X-axis minus direction (S960), and then the process proceeds to S930. For example, when the pixel position of the edge point at the end of the X axis minus direction of the inspection block is (X, Y) = (X0, Y0), the X coordinate value is from X = X0-8 to X = X0-64. After a set of a total of eight edge points that take values in the range is set as a new inspection block, the process proceeds to S930.

  On the other hand, if it is determined that the difference δY is equal to or greater than the threshold value (No in S950), the CPU 70 determines the accuracy of the edge point located in the X-axis minus direction from the X-axis minus direction end of the current inspection block as the document edge. The process moves to S970, and all the coordinate data of the edge points positioned in the X-axis minus direction from the end of the X-axis minus direction of the current inspection block are deleted from the process target data, and the process target data is Update.

  For example, if the pixel position of the edge point at the end of the current X-axis minus direction of the inspection block is (X, Y) = (X0, Y0), all the coordinate data of the edge points whose X coordinate values are less than X0 Delete the processing target data and update the processing target data. Thereafter, the second edge determination process ends. In the second edge determination process, the coordinate data having low accuracy as the document edge is deleted from the coordinate data of the edge points constituting the processing target data.

  That is, in S890 of the lower edge detection process, the coordinate data of the edge point with low accuracy is deleted from the document lower edge data as the lower edge in this way, and the document lower edge data is determined. In S890, as described above, an edge point having a larger X coordinate value is treated as having a higher probability that the edge point is an edge point (lower edge) representing the lower end of the document, and an edge point group having a larger X coordinate value. The reason why the edge point group having a small X coordinate value existing at a position greatly deviated in the Y axis direction from the document is deleted from the document lower edge data is that the inspection positions (Xe, Ye in order from Xe = 0 in the lower edge detection process) ) To search for an edge point representing the bottom edge of the document. If the document is tilted to the left, the edge point corresponding to the ruled line drawn inside the document is not the bottom edge of the document. This is because there is a possibility of erroneous detection as an edge point representing.

  Such an erroneous detection is highly likely to occur in an edge point group having a small X coordinate value, and is unlikely to occur in an edge point group having a large X coordinate value. For this reason, in S890, an edge point with a larger X coordinate value is treated as having a higher probability that the edge point is an edge point (lower edge) representing the lower end of the document, and an edge point with lower accuracy is treated as the lower edge of the document. The data is deleted from the data, and the document lower edge data is determined.

  When the processing in S890 is completed in this manner, the CPU 70 ends the lower edge detection processing in S460 and proceeds to S465. In S465, the document lower edge data determined in the lower edge detection processing is determined. Each point shown is approximated by a straight line, and an approximate straight line of the point group indicated by the document lower edge data (hereinafter referred to as a “lower edge approximate straight line”) is calculated. Thereafter, the process proceeds to S470. However, if there is no coordinate data registered as the document lower edge data in the immediately preceding lower edge detection process, it is determined that the lower edge detection has failed, and the process proceeds to S470 without calculating the lower edge approximate straight line. It shall be.

  In step S470, the CPU 70 determines the right edge detection process (S420), the left edge detection process (S450), and the lower edge detection process from the mutual relationship between the right edge approximate line, the left edge approximate line, and the lower edge approximate line. In (S460), it is determined whether or not the document edge is correctly detected. That is, it is determined whether all of the right edge, the left edge, and the lower edge have been detected normally.

  Specifically, the right edge approximate line and the left edge approximate line are parallel, the right edge approximate line and the lower edge approximate line are orthogonal, and the left edge approximate line and the lower edge approximate line are orthogonal If there is a relationship, it is determined that the document edge is correctly detected, and the determination in S470 is Yes.

  On the other hand, if detection of any one of the right edge, the left edge, and the lower edge fails, or if the relationship between the right edge, the left edge, and the lower edge is not correct, it is determined No in S470. . That is, the right edge approximate line and the left edge approximate line are parallel, the right edge approximate line and the lower edge approximate line are orthogonal, and the left edge approximate line and the lower edge approximate line are orthogonal. Except in the case of No., it is determined that the document edge is not correctly detected, and No is determined in S470. However, the determination of the mutual relationship here should be executed in consideration of an error.

If it is determined that the document edge is not correctly detected (No in S470), the CPU 70 proceeds to S471 and executes error processing similar to the processing in S431.
On the other hand, if it is determined in S470 that the document edge is correctly detected (Yes in S470), the CPU 70 proceeds to S480, and the lower left corner position of the document is determined by the intersection of the left edge approximate line and the lower edge approximate line. Assuming that there is a coordinate value, the coordinate value is calculated. Further, the lower right corner position of the document is estimated to be the intersection of the right edge approximate line and the lower edge approximate line, and the coordinate value is calculated (S485). . FIG. 13 is an explanatory diagram showing a method for estimating the document lower left corner position Ka and the document lower right corner position Kb in S480 and S485. In the figure, white circles represent edge points, and black circles represent the corner positions Ka and Kb.

  When the process in S485 is completed, the CPU 70 assumes that the document placed on the document table 11a is a standard sheet based on the position coordinates of the angular position estimated in S480 and S485. The size of the document placed on the table 11a and the document placement area on the document table 11a are estimated, and the inclination angle θ of the document placed on the document table 11a is estimated based on the right-edge approximate straight line ( S490).

  That is, the length from the lower left corner position to the lower right corner position of the document estimated in S480 and S485 is estimated as the short side length (horizontal width) of the document, and √2 times the short side length is set to the document length. Assuming that the document size is estimated by estimating the side length (vertical width), the document of the document size is placed on the document table 11a starting from the estimated lower left corner position and lower right corner position. The coordinates where the upper left corner and the upper right corner of the document are located are derived, and an area within a quadrilateral formed by connecting these four corners is estimated as the document placement area.

  In addition, in S490, the angle formed by the right edge approximate line with respect to the Y axis is estimated as the document inclination angle θ. When this process is completed, the CPU 70 ends the document estimation process in S220, proceeds to S230, and executes the subsequent process. That is, based on the estimation result in the document estimation process, a scaling factor and an inclination correction amount are set (S251, S253), and a reading area is further set (S255). Thereafter, the process proceeds to the main scan (S270).

  In the main scan (S270), the reading unit 15 is transported under the set reading area, and the reading unit 15 is caused to read an image shown in the area. Then, the read result is enlarged / reduced in accordance with the set variable magnification, and rotated in accordance with the set inclination correction amount to generate image data for printing. Also, by printing this as print target data, a copy image of the original, which has been automatically scaled and further tilt-corrected, is printed on recording paper through the printing unit 30.

  The configuration of the digital multi-function peripheral 1 according to the present embodiment has been described above. According to the multi-function peripheral 1 according to the present embodiment, a part of the readable area R0 is pre-scanned, and the document edge is based on the read result. Is detected, and the state of the original (original size, original inclination, original placement area) is estimated, so that the press can be performed at a higher speed than the conventional method of pre-scanning the entire readable area R0. The can operation can be terminated, and the state of the document can be estimated efficiently.

  In this embodiment, when the right edge and the left edge are detected, the document edge candidate point at the center of the document table 11a is treated as having a higher probability of being the document edge, and the document table 11a The document right edge data and document left edge data are determined based on the block closer to the center (the block with the largest Y coordinate value), and the right edge approximate straight line and the left edge approximate straight line are obtained based on these data. did.

  Therefore, according to the present embodiment, a ruled line or the like drawn inside the document is erroneously detected as a document edge (right edge or left edge) due to the document being inclined, and as a result, an error occurs in the approximate straight line. Can be suppressed from appearing greatly. Therefore, according to the present embodiment, it is possible to estimate the document size, the inclination angle of the document, and the document placement area with high accuracy with a pre-scan of a partial region.

  Further, in this embodiment, when the approximate straight line of the lower edge is obtained, the lower edge candidate point (edge point) on the right side of the document table 11a is treated as having a high probability of being truly the lower edge, and the original document is processed. The document lower edge data is determined on the basis of the block on the right side of the table 11a, and the lower edge approximate straight line is obtained based on these data. Therefore, according to the present embodiment, it is possible to suppress erroneous detection of the lower edge when the document is placed on the left side as much as possible, and with high accuracy, the document size, the inclination angle of the document, and the placement of the document. The placement area can be estimated.

When the document is inclined to the right side, the lower edge candidate point (edge point) on the left side of the document table 11a is more likely to be a truly lower edge. The contents of the second edge determination process may be switched according to the above. That is, when it is estimated that the document is inclined to the right side, the inspection block is sequentially switched from the block including the edge point with the smallest X coordinate value to the X axis plus direction, and closer to the left end of the document table 11a. Alternatively, the document lower edge data may be determined based on the blocks in the document, and the lower edge approximate straight line may be obtained based on these data.
[First modification]
Subsequently, a first modification will be described. However, the multifunctional machine of the first modification is such that the contents of the first edge determination process executed in S590 and S790 and the contents of the second edge determination process executed in S890 are partially changed. Since it is the same as the MFP 1 of the above embodiment, the description of the parts having the same configuration as the above embodiment will be omitted as appropriate.

FIG. 14 is a flowchart showing a first edge determination process executed by the CPU 70 in the first modification.
When starting the first edge determination process shown in FIG. 14 in S590 or S790, the CPU 70 first initializes the variable L and the variable LMAX to a value of zero (S1010). Thereafter, the process proceeds to S1015, and it is determined whether or not the preset processing target data includes coordinate data of at least eight edge points.

  If it is determined that the processing target data includes coordinate data of eight or more edge points (Yes in S1015), the process proceeds to S1020, and the processing target data does not include coordinate data of eight or more edge points. (No in S1015), the first edge determination process is terminated without executing the processes of S1020 to S1090.

  On the other hand, when proceeding to S1020, the CPU 70 sets, as the inspection block, a block that includes an edge point having the maximum Y coordinate value as an element in the processing target data, as in the process in S620. Thereafter, the process proceeds to S1030.

  In S1030, the CPU 70 updates the variable L to a value obtained by adding 8 (L ← L + 8). The variable L is a variable for measuring the length of successive edge points, and the value 8 is added here because the eight edge points constituting the inspection block are continuous. .

  When the processing in S1030 is completed in this way, the CPU 70 proceeds to S1040, and the coordinate data of the edge point adjacent in the Y-axis minus direction from the inspection block is included in the processing target data as in the processing in S630. It is determined whether or not. That is, if the pixel position of the edge point having the smallest Y coordinate value among the edge points constituting the inspection block is (X, Y) = (X0, Y0), the edge points adjacent in the Y axis minus direction are As the coordinate data, it is determined whether or not the coordinate data of the edge point whose Y coordinate value is Y = Y0-8 is included in the processing target data.

  If it is determined that the coordinate data of the edge point adjacent in the Y axis minus direction is included in the processing target data (Yes in S1040), the process proceeds to S1050, and the coordinate data of the edge point adjacent in the Y axis minus direction is obtained. If it is determined that the data is not included in the processing target data (No in S1040), the process proceeds to S1071.

  In S1050, the CPU 70, like the process in S640, determines the X coordinate value of the edge point at the end of the Y-axis minus direction of the inspection block and the X of the edge point adjacent to the inspection block in the Y-axis minus direction. A difference δX from the coordinate value is calculated. That is, the pixel position of the edge point at the end of the inspection block in the Y-axis minus direction is (X, Y) = (X0, Y0), and the pixel position of the edge point adjacent to the inspection block in the Y-axis minus direction is (X, Y). When Y) = (X1, Y0-8), δX = | X0−X1 | is calculated.

  When this process is finished, the CPU 70 determines whether or not the difference δX is less than a threshold value determined in advance in the design stage (S1055). If the CPU 70 determines that the difference δX is less than the threshold value (Yes in S1055), The block is updated to the next block adjacent in the Y-axis minus direction (S1060), and then the process proceeds to S1030. For example, when the pixel position of the edge point at the end of the Y-axis minus direction of the inspection block is (X, Y) = (X0, Y0), the Y coordinate value ranges from Y = Y0-8 to Y = Y0-64. After a set of a total of eight edge points that take values in the range is set as a new inspection block, the process proceeds to S1030.

  On the other hand, when determining that the difference δX is equal to or greater than the threshold (No in S1055), the CPU 70 proceeds to S1071. In S1071, it is determined whether or not the current value of the variable L is greater than the variable LMAX. If it is determined that the current value of the variable L is greater than the variable LMAX (Yes in S1071), finally the variable L = 0. The edge points that make up each block set as the inspection block from the time of setting (initialization) to the present are determined to be the longest continuous edge point set, and the coordinate data of the edge points belonging to this set are determined. The final candidate data is recorded in the RAM 80 (S1073). Here, the last candidate data recorded last time is overwritten, and the latest last candidate data is recorded in the RAM 80, thereby discarding the previous value of the last candidate data.

  When the process in S1073 is completed, the CPU 70 proceeds to S1075 and updates the value of the variable LMAX to the current value of the variable L (LMAX ← L). Thereafter, the process proceeds to S1080.

  On the other hand, if it is determined in S1071 that the current value of the variable L is less than or equal to the value of the variable LMAX (No in S1071), the CPU 70 proceeds to S1080 without executing the processing of S1073 and S1075.

  In S1080, the CPU 70 determines whether or not an edge point located in the Y-axis minus direction from the Y-axis minus direction end of the inspection block is registered in the processing target data. That is, when the pixel position of the edge point at the end in the Y-axis minus direction of the current inspection block is (X, Y) = (X0, Y0), an edge point having a Y coordinate value less than Y0 is registered in the processing target data. Judge whether or not.

  If it is determined that an edge point located in the Y-axis minus direction from the Y-axis minus direction end of the inspection block is registered in the processing target data (Yes in S1080), the process proceeds to S1081, and the value of the variable L is set to zero. (L = 0). Thereafter, the inspection block is updated to the next block located in the negative Y-axis direction from the current inspection block (S1085), and the process proceeds to S1030.

  For example, when the pixel position of the edge point at the end of the Y-axis minus direction of the inspection block is (X, Y) = (X0, Y0), among the edge points whose Y coordinate value is less than Y = Y0, Y Starting from the edge point with the largest coordinate value, a set of a total of eight edge points in the negative Y-axis direction from this edge point is set as a new inspection block (S1085), and the process proceeds to S1030.

  In this way, the CPU 70 repeatedly executes the processing of S1030 to S1085, and when the edge point located in the Y-axis minus direction from the end of the Y-axis minus direction of the inspection block is not in the processing target data (No in S1080). At that time, the coordinate data of each edge point other than the edge points registered in the RAM 80 as final candidate data is deleted from the processing target data, and the processing target data is updated. Thereafter, the first edge determination process ends.

  In the first edge determination process, a set of edge points that are longest and continuous in the Y-axis direction is treated as a highly accurate edge point as a document edge in this way, and edge points other than the edge points belonging to this set are handled. By deleting the coordinate data from the processing target data, the coordinate data having low accuracy as the document edge is deleted from the processing target data.

FIG. 15 is an explanatory diagram showing a method for determining document right edge data in the first modification.
In the example illustrated in FIG. 15, the lengths of sets of continuous edge points measured as the value of the variable L are L [0], L [1], and L [2], respectively, and the magnitude relationship is L [0]. Since 1]> L [2]> L [0], the set of edge points of length L [1] is treated as a set of edge points with high accuracy as the right edge, and the edges belonging to this set Only the point coordinate data is left in the document right edge data, and the document right edge data is determined. Finally, a right edge approximate straight line is obtained from the set of edge points.

According to such a procedure, in the first modification, the document right edge data is determined in S590, and the document left edge data is determined in S790.
On the other hand, in S890, the document lower edge data is set as processing target data, and the second edge determination process shown in FIG. 16 is executed to determine the document lower edge data. FIG. 16 is a flowchart showing the second edge determination process executed by the CPU 70 in the first modification.

  When the second edge determination process shown in FIG. 16 is started, the CPU 70 first initializes the variable L and the variable LMAX to a value of zero (S1110). Thereafter, the process proceeds to S1115, and it is determined whether or not the preset processing target data includes coordinate data of at least eight edge points.

  If it is determined that the processing target data includes coordinate data of eight or more edge points (Yes in S1115), the process proceeds to S1120, and the processing target data does not include coordinate data of eight or more edge points. (No in S1115), the second edge determination process is terminated without executing the processes in S1120 to S1190.

  On the other hand, when the process proceeds to S1120, the CPU 70 sets, as the inspection block, a block that includes an edge point having the maximum X coordinate value in the processing target data as in the process in S920. Thereafter, the process proceeds to S1130.

  When the process proceeds to S1130, the CPU 70 updates the variable L to a value obtained by adding 8 (L ← L + 8), and then proceeds to S1140. Similar to the process in S930, the CPU 70 is adjacent to the inspection block in the X axis minus direction. It is determined whether the coordinate data of the edge point is included in the processing target data. That is, if the pixel position of the edge point having the smallest X coordinate value among the edge points constituting the inspection block is (X, Y) = (X0, Y0), the edge points adjacent in the X axis minus direction As the coordinate data, it is determined whether or not the coordinate data of the edge point whose X coordinate value is X = X0-8 is included in the processing target data.

  If it is determined that the coordinate data of the edge point adjacent in the X axis minus direction is included in the processing target data (Yes in S1140), the process proceeds to S1150, and the coordinate data of the edge point adjacent in the X axis minus direction is obtained. If it is determined that the data is not included in the processing target data (No in S1140), the process proceeds to S1171.

  In step S1150, as in the process in step S940, the CPU 70 determines the X coordinate value of the edge point at the end of the X axis minus direction of the inspection block and the X of the edge point adjacent to the inspection block in the X axis minus direction. A difference δY from the coordinate value is calculated. That is, the pixel position of the edge point at the end of the inspection block in the X-axis minus direction is (X, Y) = (X0, Y0), and the pixel position of the edge point adjacent to the inspection block in the X-axis minus direction is (X, Y). When Y) = (X0-8, Y1), δY = | Y0−Y1 | is calculated.

  When this process is finished, the CPU 70 determines whether or not the difference δY is less than a threshold value determined in advance in the design stage (S1155). If the CPU 70 determines that the difference δY is less than the threshold value (Yes in S1155), the inspection is performed. The block is updated to the next block adjacent in the X-axis minus direction (S1160), and then the process proceeds to S1130. For example, when the pixel position of the edge point at the end of the X axis minus direction of the inspection block is (X, Y) = (X0, Y0), the X coordinate value is from X = X0-8 to X = X0-64. After a set of a total of eight edge points that take values in the range is set as a new inspection block, the process proceeds to S1130.

  On the other hand, when determining that the difference δY is equal to or greater than the threshold (No in S1155), the CPU 70 proceeds to S1171. In S1171, it is determined whether or not the current value of the variable L is greater than the variable LMAX. If it is determined that the current value of the variable L is greater than the variable LMAX (Yes in S1171), finally the variable L = 0. It is determined that the edge points constituting each block set as the inspection block from the time of setting (initialization) to the present are the set of the longest continuous edge points, and this set is included in the set in the same manner as in S1073. The coordinate data of the edge point to which it belongs is recorded in the RAM 80 as final candidate data (S1173).

  When the processing in S1173 is completed, the CPU 70 proceeds to S1175 and updates the value of the variable LMAX to the current value of the variable L (LMAX ← L). Thereafter, the process proceeds to S1180.

  On the other hand, if it is determined in S1171 that the current value of the variable L is equal to or less than the value of the variable LMAX (No in S1171), the CPU 70 proceeds to S1180 without executing the processes of S1173 and S1175.

  In S1180, the CPU 70 determines whether an edge point located in the X-axis minus direction from the X-axis minus direction end of the inspection block is registered in the processing target data. That is, when the pixel position of the edge point at the end of the X axis minus direction of the current inspection block is (X, Y) = (X0, Y0), an edge point having an X coordinate value less than X0 is registered in the processing target data. Judge whether or not.

  If it is determined that the edge point located in the X-axis minus direction from the X-axis minus direction end of the inspection block is registered in the processing target data (Yes in S1180), the process proceeds to S1181, and the value of the variable L is set to zero. (L = 0). Thereafter, the inspection block is updated to the next block positioned in the minus direction of the X axis from the current inspection block (S1185), and the process proceeds to S1130.

  For example, when the pixel position of the edge point at the end of the X-axis minus direction of the inspection block is (X, Y) = (X0, Y0), among the edge points whose X coordinate value is less than X = X0, X Starting from the edge point with the maximum coordinate value, a set of a total of eight edge points in the negative X-axis direction from this edge point is set as a new inspection block (S1185), and the process proceeds to S1130.

  In this way, the CPU 70 repeatedly executes the processing of S1130 to S1185, and when the edge point located in the X-axis minus direction from the end of the X-axis minus direction of the inspection block is not in the processing target data (No in S1180). At that time, the coordinate data of each edge point other than the edge points registered in the RAM 80 as final candidate data is deleted from the processing target data, and the processing target data is updated. Thereafter, the second edge determination process ends.

  In the second edge determination process, a set of edge points that are longest in the X-axis direction is treated as a highly accurate edge point as a document edge in this way, and edge points other than the edge points belonging to this set are handled. By deleting the coordinate data from the processing target data, the coordinate data having low accuracy as the document edge is deleted from the processing target data.

  In this way, in the first modification, only the coordinate data of the edge points belonging to the set of edge points that are longest in the X-axis direction are left in the document lower edge data, and the document lower edge data is determined. Finally, a lower edge approximate straight line is obtained from the set of edge points.

  The first modification has been described above. In the first modification, the longest continuous edge point set is treated as the most accurate edge point set as the document edge, and the right edge approximate straight line and the left edge approximate straight line are handled. And a lower edge approximate straight line.

  Therefore, according to the first modification, it is possible to prevent as much as possible from obtaining an approximate straight line as an approximate straight line of the document edge by being confused by ruled lines or dust inside the original, Thus, it is possible to estimate the document size, the inclination angle of the document, and the document placement area with high accuracy.

[Second modification]
Subsequently, a second modification will be described. However, the MFP of the second modified example is such that the contents of the first edge determination process executed in S590 and S790 and the contents of the second edge determination process executed in S890 are partially changed. Since it is the same as the multifunction device 1 of the above-described embodiment, the description of the parts having the same configuration as the above-described embodiment will be appropriately omitted below.

FIG. 17 is a flowchart showing a first edge determination process executed by the CPU 70 in the second modification.
When the CPU 70 starts the first edge determination process shown in FIG. 17 in S590 or S790, first, in S1210, whether the preset processing target data includes coordinate data of at least eight or more edge points. Judge whether or not.

  If it is determined that the processing target data includes coordinate data of eight or more edge points (Yes in S1210), the process proceeds to S1220, and the processing target data does not include coordinate data of eight or more edge points. (No in S1210), the first edge determination process is terminated without executing the processes in S1220 to S1290.

  On the other hand, when the process proceeds to S1220, the CPU 70 sets, as the inspection block, a block that includes an edge point having the maximum Y coordinate value as an element in the processing target data, similarly to the process in S620. Thereafter, the process proceeds to S1230.

  In step S1230, the CPU 70 calculates approximate straight lines of the edge points based on the coordinate data of a total of eight edge points belonging to the inspection block. Then, the calculated inclination and intercept values of the approximate line are stored in the RAM 80 in association with the coordinate data of the edge point used for calculating the approximate line (S1240). Thereafter, the process proceeds to S1250.

  In S1250, the CPU 70 determines whether or not an edge point located in the Y-axis minus direction from the Y-axis minus direction end of the inspection block is registered in the processing target data. That is, when the pixel position of the edge point at the end in the Y-axis minus direction of the current inspection block is (X, Y) = (X0, Y0), an edge point having a Y coordinate value less than Y0 is registered in the processing target data. Judge whether or not.

  If it is determined that the edge point located in the Y-axis minus direction from the Y-axis minus direction end of the inspection block is registered in the processing target data (Yes in S1250), the inspection block is set to be more than the current inspection block. The next block located in the negative axis direction is updated (S1260), and then the process proceeds to S1230.

  For example, when the pixel position of the edge point at the end of the Y-axis minus direction of the inspection block is (X, Y) = (X0, Y0), among the edge points whose Y coordinate value is less than Y = Y0, Y Starting from the edge point with the largest coordinate value, a set of a total of eight edge points in the negative Y-axis direction from this edge point is set as a new inspection block (S1260), and the process proceeds to S1230.

  In this way, the CPU 70 repeatedly executes the processing of S1320 to S1260, obtains the approximate straight line of the edge point registered in the processing target data for each block, and calculates the slope and intercept value of the approximate straight line of each block. Then, it is stored in the RAM 80 in association with the coordinate data of the edge point used for the calculation of the approximate straight line.

  Then, when the edge point located in the Y-axis minus direction from the Y-axis minus direction end of the inspection block is not in the processing target data (No in S1250), from the inclination of the approximate straight line of each block accumulated in the RAM 80 until then. Histogram data about the slope is generated (S1270). FIG. 18 is an explanatory diagram showing the histogram data of the slope generated in S1270 as a graph.

  In this modification, a plurality of classes are set for the inclination by dividing the range of inclination that the approximate straight line can take into a plurality of classes. In S1270, the slope of each block is classified into classes corresponding to the values, and the histogram data of the slope is generated by calculating the frequency of each class.

  Then, the class having the highest frequency is specified from the histogram data of the slope, and all the blocks that have been used to calculate the approximate straight line of the slope belonging to the class having the highest frequency are selected from the processing target data ( S1275).

  When the processing in S1275 is finished, the CPU 70 generates histogram data for the intercept of the approximate line of each block selected in S1275 based on the information of the intercept of the approximate line of each block accumulated in the RAM 80 (S1280). ).

  That is, from the information on the intercept of the approximate straight line of each block stored in the RAM 80, the information on the intercept for the block selected in S1275 is selectively used to generate histogram data for the intercept.

  In this modification, a plurality of classes are set for the intercept by dividing the range of the intercept that the approximate straight line can take into a plurality of classes. In S1280, the intercept of each block selected in S1275 is classified into classes according to the values, and the frequency of each class is calculated, thereby generating intercept histogram data.

  Then, from the histogram data of the intercept, the class having the highest frequency for the intercept is identified, and all of the blocks that have been used to calculate the approximate straight line of the intercept belonging to the identified class are selected from the block group selected in S1275. (S1285).

  After completing the processing in S1285, the CPU 70 proceeds to S1290, deletes the coordinate data of the edge points other than the edge point belonging to the block selected in S1285 from the processing target data, and stores the processing target data in the processing target data. The processing target data is updated so that only the coordinate data of the edge points belonging to the block selected in S1285 remains. Thereafter, the first edge determination process ends.

  In this way, in the first edge determination process of the second modified example, the set of blocks for which the approximate straight line having the highest frequency is calculated is treated as a set of edge points with high accuracy that are document edges, and these The processing target data is updated so that only the coordinate data of the edge point remains.

According to such a procedure, in the second modification, the document right edge data is determined in S590, and the document left edge data is determined in S790.
On the other hand, in S890, the document lower edge data is set as processing target data, and the second edge determination process shown in FIG. 19 is executed to determine the document lower edge data. FIG. 19 is a flowchart illustrating a second edge determination process executed by the CPU 70 in the second modification.

  When the second edge determination process shown in FIG. 19 is started, the CPU 70 first determines in S1310 whether the preset processing target data includes coordinate data of at least eight or more edge points. .

  If it is determined that the processing target data includes coordinate data of eight or more edge points (Yes in S1310), the process proceeds to S1320, and the processing target data does not include coordinate data of eight or more edge points. (No in S1310), the second edge determination process is terminated without executing the processes in S1320 to S1390.

  On the other hand, when proceeding to S1320, the CPU 70 sets, as the inspection block, a block that includes an edge point having the maximum X coordinate value as an element in the processing target data, as in the processing in S920. Thereafter, the process proceeds to S1330.

  In step S1330, the CPU 70 calculates approximate straight lines of the edge points based on the coordinate data of a total of eight edge points belonging to the inspection block. Then, the calculated inclination and intercept values of the approximate line are stored in the RAM 80 in association with the coordinate data of the edge point used for calculating the approximate line (S1340). Thereafter, the process proceeds to S1350.

  In S1350, the CPU 70 determines whether an edge point located in the X-axis minus direction from the X-axis minus direction end of the inspection block is registered in the processing target data. That is, when the pixel position of the edge point at the end of the X axis minus direction of the current inspection block is (X, Y) = (X0, Y0), an edge point having an X coordinate value less than X0 is registered in the processing target data. Judge whether or not.

  If it is determined that the edge point located in the X-axis minus direction from the X-axis minus direction end of the inspection block is registered in the processing target data (Yes in S1350), the inspection block is set to be more X than the current inspection block. The block is updated to the next block located in the negative axis direction (S1360), and thereafter, the process proceeds to S1330.

  For example, when the pixel position of the edge point at the end of the direction of the inspection block is (X, Y) = (X0, Y0), among the edge points whose X coordinate value is less than X = X0, the X coordinate value is Starting from the maximum edge point, a set of a total of eight edge points in the negative X-axis direction from this edge point is set as a new inspection block (S1360), and then the process proceeds to S1330.

  In this way, the CPU 70 repeatedly executes the processes of S1330 to S1360, obtains the approximate straight line of the edge point registered in the processing target data for each block, and calculates the slope and intercept value of the approximate straight line of each block. Then, it is stored in the RAM 80 in association with the coordinate data of the edge point used for the calculation of the approximate straight line.

  Then, when the edge point located in the X-axis minus direction from the X-axis minus direction end of the inspection block is not in the processing target data (No in S1350), from the inclination of the approximate straight line of each block accumulated in the RAM 80 until then. In the same manner as the processing in S1270, histogram data for the slope is generated (S1370). Thereafter, the class having the highest frequency with respect to the slope is identified from the histogram data of the slope, and all of the blocks that have been used to calculate the approximate straight line of the slope belonging to the class having the highest frequency are selected from the processing target data. (S1375).

  When the processing in S1375 is completed, the CPU 70 is information on the intercept of the approximate line stored in the RAM 80, and is similar to the process in S1280 based on the information on the intercept of the approximate line of each block selected in S1375. Then, histogram data for these intercepts is generated (S1380). That is, in S1380, the intercept of each block selected in S1375 is classified into classes according to the values, and the frequency of each class is calculated, thereby generating intercept histogram data.

  Then, from the histogram data of the intercept, the class having the highest frequency for the intercept is identified, and all of the blocks that have been used to calculate the approximate straight line of the intercept belonging to the identified class are selected from the block group selected in S1375. Select (S1385).

  When the processing in S1385 is finished, the CPU 70 proceeds to S1390, deletes the coordinate data of edge points other than the edge point belonging to the block selected in S1385 from the processing target data, and stores the processing target data in the processing target data. The processing target data is updated so that only the coordinate data of the edge points belonging to the block selected in S1385 remains. Thereafter, the second edge determination process ends.

  In this way, in the second edge determination process of the second modification, the set of blocks for which the approximate straight line having the highest frequency is calculated is the set of highly accurate edge points that are document edges (lower edges). The processing target data is updated so that only the coordinate data of these edge points remains. According to such a procedure, in the second modified example, the document lower edge data is determined in S890.

  The second modification has been described above. In the second modification, each edge point of the block group for which the approximate straight line having the highest frequency is calculated is treated as a highly accurate edge point as a document edge, and right edge approximation is performed. A straight line, a left edge approximate line, and a lower edge approximate line are obtained.

Therefore, according to the second modified example, it is possible to prevent as much as possible the determination of an incorrect approximate line as an approximate line of the document edge, which is confused by ruled lines, dust, etc. inside the document, It is possible to estimate the document size, the document inclination angle, and the document placement area with a high degree of accuracy by pre-scanning.
[Correspondence with Claims]
Although the embodiments of the present invention have been described above, the image analysis function of the image reading apparatus of the present invention is realized by the CPU 70 as software. The reading control unit is realized by the reading control unit 20.

  The conversion means is realized by the processing of S215 executed by the CPU 70, and the selection means is realized by the processing of S510 to S580, S710 to S780, S810 to S880 executed by the CPU 70. In particular, the operation of “detecting an edge point to be inspected for continuity” is realized by processing such as S525, S725, and S825.

  In addition, the detection means is realized by the processing of S590, S790, and S890 executed by the CPU. In particular, the inspection object switching setting means is realized by an operation of switching and setting the inspection block. In the above-described embodiment, image data obtained by pre-scanning a partial area is set as inspection target data to detect a document edge, and the pre-scan area is located on the end side from the center of the document table. Therefore, in this embodiment, in the first edge determination process, by setting a block including an edge point having the maximum Y coordinate value as an inspection block, the group closest to the center of the document table is set as the inspection target. Set to block.

Further, the first determination means is realized by processing such as S630, S930, S1040, and S1140, the divergence degree calculating means is realized by processing such as S640, S940, S1050, and S1150, and the second determination means is S650, This is realized by the processing of S950, S1055, S1155 and the like.
[Others]
Further, the present invention is not limited to the above-described embodiments, and can take various forms. For example, in the above-described embodiment, the example in which the present invention is applied to the digital multifunction peripheral 1 including the contact image sensor as a reading unit has been described. However, the present invention is limited to an image reading apparatus including the contact image sensor as a reading unit. Needless to say, it is not something.

1 is a block diagram illustrating a configuration of a digital multifunction machine 1. FIG. It is explanatory drawing which showed readable area | region R0 of the reading glass. 6 is a flowchart illustrating a copy control process executed by a CPU. 6 is a flowchart showing an automatic scaling copying process executed by a CPU. 6 is a flowchart showing document estimation processing executed by CPU. It is a flowchart showing the right edge detection process which CPU70 performs. It is explanatory drawing which showed the structure of edge image data. It is a flowchart showing the 1st edge determination process which CPU70 performs. FIG. 6 is an explanatory diagram illustrating a method for determining original right edge data. It is a flowchart showing the left edge detection process which CPU70 performs. It is a flowchart showing the lower edge detection process which CPU70 performs. It is a flowchart showing the 2nd edge determination process which CPU70 performs. FIG. 5 is an explanatory diagram showing a method for estimating a document lower left corner position and a document lower right corner position. It is a flowchart showing the 1st edge determination process of a 1st modification. It is explanatory drawing showing the determination method of the original right edge data in a 1st modification. It is a flowchart showing the 2nd edge confirmation process of a 1st modification. It is a flowchart showing the 1st edge determination process of a 2nd modification. It is explanatory drawing showing histogram data. It is a flowchart showing the 2nd edge confirmation process of a 2nd modification.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Digital compound machine, 10 ... Image reading part, 11 ... Reading glass, 11a ... Original plate, 13 ... Housing | casing, 13a ... Frame, 15 ... Reading unit, 16 ... Conveyance mechanism, 17 ... Motor, 20 ... Reading control part , 30 ... printing section, 40 ... print control section, 50 ... display operation section, 60 ... communication section, 70 ... CPU, 80 ... RAM, 90 ... flash memory, MK ... mark, P ... original, R0 ... readable area

Claims (6)

A document table on which a document to be read is placed;
A reading unit for reading a document placed on the document table;
Reading control means for generating image data representing a read image of the document placed on the document table by causing the reading unit to read the document placed on the document table ;
The image data generated by the reading control means, conversion means for converting the image data representing an edge image,
The candidate edge points representing the edge of the document and extending in a specific direction, selecting means for selecting from the image data converted by said converting means,
Storage means for storing position information of the selected edge point as the candidate by the selection means;
Based on the distribution of the edge points selected as the candidates by the selection means specified from the position information of the edge points stored in the storage means , the original from the group of candidates selected by the selection means By discarding edge points with low accuracy as edge points representing the edges of the document, the candidates remaining after the discard are detected as edge points representing the edges of the document extending in the specific direction in the converted image data Detecting means for
With
The selection means includes
The end point of the converted image data is set as a reference start position, and the converted image data is moved from the reference start position to the inside of the image data along a predetermined first direction. Referring to the pixel value of each pixel that constitutes, the edge point to be inspected for continuity is detected, and the detected edge point is used as a starting point in a second direction perpendicular to the first direction. , By referring to the pixel value of each pixel constituting the image data after the conversion and checking the continuity of the edge points in the second direction, a predetermined amount of edges in the second direction points, as candidate edge points representing the edge of the document and extending in the specific direction, the operation of selecting the image data after the conversion, a plurality of times repeatedly performed while shifting the reference start position in the second direction By specifying Candidate edge points representing the edge of the document extending direction, by a plurality of the predetermined amount, the image reading apparatus characterized by being in the configuration of selecting in search from the end of the image data. The reading control unit has a function of causing the reading unit to read an area shorter than an area extending from one side of the document table perpendicular to the second direction to a center extending in the second direction;
The selection means selects the edge point candidates from the image data converted by the conversion means for the image data generated by the reading control means by the function,
The detection means sets the predetermined amount of candidates selected by the selection means as one group unit, and determines the continuity of each group in the second direction in the second direction closest to the center of the document table. An edge point that represents the edge of the document extending in the specific direction by discarding all of the candidates belonging to a group that is discontinuous from the group that is a criterion as a criterion and discarding all of the candidates as the low-accuracy candidates The image reading apparatus according to claim 1, wherein: The detection means includes
Inspection object switching setting means for setting the inspection target group so as to sequentially switch from the group closest to the center of the document table to the group closest to the edge in the second direction;
In the direction in which the inspection target group is switched by the inspection target switching setting unit, whether or not there is a group adjacent to the inspection target group currently set by the inspection target switching setting unit is determined in the second direction. First determination means for evaluation and determination in a one-dimensional space parallel to
When the first determination unit determines that the adjacent group exists, the candidate belonging to the group to be inspected and closest to the adjacent group and the candidate belonging to the adjacent group A deviation degree calculating means for calculating a deviation degree in the first direction of the candidate closest to the group to be inspected;
Second determination means for determining whether or not the deviation degree calculated by the deviation degree calculation means is less than a threshold;
And when the first determination means determines that the adjacent group does not exist or the second determination means determines that the degree of divergence is equal to or greater than a threshold, the group to be inspected up to that point All of the groups located on the edge side of the document table in the second direction that are not set to be considered as discontinuous groups from the group closest to the center of the document table in the second direction, The image reading apparatus according to claim 2, wherein all of the candidates belonging to the discontinuous group are configured to be discarded. The detection means evaluates the continuity of each group in the second direction with the predetermined amount of candidates selected by the selection means as one group unit, and sets a set of longest consecutive groups as the accuracy. The candidate belonging to the group that does not belong to the longest consecutive group set is regarded as the low-accuracy candidate, and the low-accuracy candidate is discarded. The image reading apparatus according to claim 1, wherein an edge point representing an edge of the document extending in the direction is detected. The detection means includes
Inspection object switching setting means for setting the inspection object group to sequentially switch in the second direction;
In the direction in which the inspection target group is switched by the inspection target switching setting unit, whether or not there is a group adjacent to the inspection target group currently set by the inspection target switching setting unit is determined in the second direction. First determination means for evaluation and determination in a one-dimensional space parallel to
When the first determination unit determines that the adjacent group exists, the candidate belonging to the group to be inspected and closest to the adjacent group and the candidate belonging to the adjacent group A deviation degree calculating means for calculating a deviation degree in the first direction of the candidate closest to the group to be inspected;
Second determination means for determining whether or not the deviation degree calculated by the deviation degree calculation means is less than a threshold;
The inspection target group determined by the first determination means that the adjacent group is present and the divergence degree is determined by the second determination means to be less than a threshold, and the group After the group, it is evaluated that the adjacent group set as the group to be inspected is continuous, and the first determination unit determines that the adjacent group does not exist, or the second determination With respect to the group to be inspected, which is determined by the means that the degree of divergence is equal to or greater than a threshold value, the group and the group set as the group to be inspected next to the group are evaluated as not consecutive. 5. The image reading apparatus according to claim 4, wherein a set of the longest continuous groups is detected. The detection means includes
The approximate line for each group is calculated using the predetermined amount of candidates selected by the selection means as a group unit, and the histogram data for the approximate line based on the calculated slope and intercept information for each group. A set of groups for which an approximate straight line with the highest frequency is calculated based on the histogram data generated by the histogram data generation unit is regarded as a set of candidates with high accuracy. The candidate belonging to a group that does not belong to the set of groups for which the approximate straight line with the highest frequency is calculated is regarded as a candidate with low accuracy, and the candidate with low accuracy is discarded, thereby extending in the specific direction. The image reading apparatus according to claim 1, wherein an edge point representing an edge of the document is detected.

Images