JPH06259598A - Method and device for inputting image - Google Patents

Abstract

(57) [Abstract] [Purpose] To provide a means for reducing the time required for total input in an image input method for inputting a plurality of image data. [Structure] For example, a scanner 400, a memory 401, a control unit 402, an encoding circuit 403, a communication circuit 404, a storage device 405, a printer 406, a frame memory 407, a display 408, a black run extraction circuit 700, and a bus 409. There is a device configured as. [Effect] When images of the same type are continuously input, the time required to extract the characteristic information of the images is shortened, and the time required for image input is significantly shortened.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of inputting an image in, for example, a facsimile apparatus, a copying machine, an electronic filing apparatus, etc., and more particularly to a means for inputting a plurality of images.

[0002]

2. Description of the Related Art Conventionally, in a copying machine or the like, when an image is placed on a document table, the image is placed at an angle with respect to a standard position.
It is an empirical fact that the tilted image data may be input.

Further, the image data that should have been placed at the standard position on the platen, that is, the image data that should have been placed straight, often tilts due to wind pressure when the document pressing plate is lowered.

In order to prevent the inclination of the image input data, the operation for inputting the image data must be carefully performed.

Such a careful work is often a burden on the operator. In particular, when the number of originals to be input is large, the burden is further increased.

In order to solve these problems, for example, the means described in Japanese Patent Laid-Open No. 62-14277 is provided.

In the above-mentioned conventional technique, the tilt angle of the image data that is tilted and input is automatically detected, and the tilt angle is corrected by rotating the input image data by the detected angle. Is.

The tilt angle detection processing in the above conventional technique will be described with reference to the drawings.

FIG. 2 shows image data that is input while being inclined.

First, the outline of a character (indicated by a rectangle in the figure)
The coordinates of the contour points P1, P2, P3, etc., which are the lower right points of, are extracted.

The most straight line on which these contour points are on the same straight line is detected, and the angle formed by the straight line from the horizontal direction is obtained, and this angle is taken as the inclination angle of the image.

Then, by the detected tilt angle,
The image is corrected by rotating the image.

As described above in the prior art,
There has been proposed an image processing unit that performs inclination correction processing on image data that is input with inclination.

[0014]

However, in the above-mentioned prior art, only the processing for one original is considered, and for example, the processing for continuously inputting a plurality of images is considered. Not not.

Therefore, when a plurality of images are input by the means of the above-mentioned conventional technique, there arises a problem that the image processing time increases.

The cause will be described below.

First, there are various types of originals input via a scanner, which is one of the image inputting means. Therefore, the inclination correction process shown in the above-mentioned conventional technique must be changed depending on the type of document.

Differences in processing due to different kinds of input images will be described below with reference to the drawings.

The inclination angle detection process shown in FIG. 2 is for horizontal characters.

Further, FIG. 3 shows a tilt detection process for vertically written characters.

As shown in FIG. 2, straight lines passing through the contour points P1, P2, P3, etc. of horizontally written characters are aligned on the straight line with a certain inclination from the horizontal direction, but in the vertical direction. In many cases, it is not available.

Further, since the inclination angle of the original at the time of scanner input is not so large, it is more efficient to search for a straight line slightly inclined from the horizontal direction in order to detect the inclination angle.

However, the straight lines passing through the contour points Q1, Q2, Q3, etc. of the vertically written characters as shown in FIG. 3 are aligned on the straight lines with a certain inclination from the vertical direction. Not aligned horizontally. Therefore, it is more efficient to search for a straight line slightly displaced from the vertical direction.

As described above, the processing is different between the vertically written characters and the horizontally written characters. Here, the tilt angle cannot be detected even if the tilt angle detection process for the horizontally written character is performed on the vertically written character image.

The type of document input from the scanner is 1
Since it is not a type, there is no choice but to adopt a method of detecting the tilt angle by performing various tilt detection processes by trial and error.

As described above, generally, in an information processing apparatus that performs some processing on input image data, processing corresponding to the type of the input image is prepared in advance, and the input image is processed. Search for the appropriate process for
You need to perform some processing. Therefore, the time required for image input will increase.

In particular, when a large number of originals are continuously input, various processes are tried on one image and one image, so that it takes a long time to input the image. There is.

Therefore, it is an object of the present invention to provide an image inputting means for reducing the inputting time of a plurality of images in an image inputting method for performing some image processing on an input image.

[0029]

In order to solve the above problems and achieve the object of the present invention, the following means can be considered.

In the image input method of optically scanning the paper surface to read image data, when inputting and reading a plurality of originals, at least one or more features relating to the image data extracted in the previous image input processing. This is an image input method for performing the image input process this time using at least one of the information. In this case, the characteristic information is
This is information that includes the direction in which the characters on the manuscript are written and the size of the characters. Based on this information, the tilt angle of the manuscript is detected, the rotation process corresponding to that angle is performed, and this time the input is made. An image input method for correcting the tilt angle of the image is also preferable.

As hardware for realizing the above method,
The following means can be considered.

An image input unit for taking in image data, a storage unit for storing the image data, and an image processing unit having a function of extracting image characteristic information including directionality from the stored image data. An image input device configured to include a correction processing unit having a function of detecting an inclination angle from a certain reference line, performing rotation processing, and correcting the inclination angle based on the image feature information. Can also be considered. Further, an image input device having an automatic paper feeding means in the image input device and having a function of continuously performing image input operation is also preferable.

An electronic filing device, a copying machine, a facsimile device, etc. equipped with this device are also conceivable.

Further, instead of using the characteristic information of the image obtained by the image input method, an image input method in which the information selected by the operator from the characteristic information preset by the operator is used as the characteristic information is also conceivable. In this case, it is preferable that the characteristic information is information including the size of characters on the original and the thickness of the line segment on the original.

In the image inputting method of optically scanning the paper surface to read the image data, a plurality of originals in which at least one of the information including the direction of the character, the size of the character, and the thickness of the line segment is the same is input. In this case, the processing used in the image input processing for the first image is omitted in the image input processing for the second and subsequent images, and compared with the time required for the image input processing for the first image, 2 An image input method in which the time required for the image input processing for the first and subsequent sheets is shortened is also conceivable.

[0036]

As described above, the image input method according to the present invention relates to a method for sequentially scanning a plurality of images and reading image data in an electronic file device, a copying machine, a facsimile device and the like.

The above-mentioned object is constituted, for example, by having an image input means for optically scanning the surface of a document and writing the image data in the memory, and a control section for executing various image processes on the image data in the memory. It is realized by an image processing device.

The image input means is realized by, for example, a scanner device. The control unit may be, for example, a CPU or RO.
It can be realized by programs stored in advance in M, RAM, various CMOSs, and ROM.

The contents of various image processing executed by the control unit include image state detection processing such as inclination angle detection processing of image data, and image correction processing such as rotation processing for an image based on the detected image state. Etc. are possible.

For the image state detection processing, a plurality of image processing methods suitable for each type of image, such as vertical writing characters and horizontal writing characters, are prepared in advance.

The control unit sequentially tries these plural existing image processing methods on the input image data, and searches for an image processing method capable of detecting the image state.

Then, based on the image state detected by the detectable image processing method, the control unit performs the image correction process to the image output means such as the printing device and the storage device.
The corrected image data is output.

Further, the control unit stores the image processing method capable of detecting the state of the input image and preferentially tries the image processing method capable of detecting the state at the next image input. In addition, it is preferable to change the image state detection procedure.

The procedure of the image input method according to the present invention will be described below.

First, the image input means stores the first image data in the memory.

Next, the control unit searches the image data on the memory for an image processing method capable of detecting the state of the image by sequentially trying the image processing for each image type, and the state of the image is searched. To detect.

Based on the detected image state, the control unit executes a correction process on the image data on the memory,
The corrected image data is output to an image output unit such as a printing device or a storage device.

With the above processing, the first image input processing is completed.

Here, in the procedure of the state detection processing for the second image data, the control unit changes so that the processing in which the state can be detected for the first image data is tried first. .

Then, the image input means stores the second image data in the memory.

Next, the control unit performs the same processing as the first sheet,
An image processing method capable of detecting the state of the image is searched for, the state of the image is detected, the image is corrected, and the corrected image data is output to an image output unit such as a printing device or a storage device.

Here, when a plurality of images are continuously input to an electronic filing device, a copying machine or the like, a set of documents or the like is input, so that each image data often has similar characteristics.

For example, when the first image is a vertically-written character image, the second and subsequent images are also likely to be vertically-written characters.

Therefore, since the image processing method that was effective for the first image is highly likely to be effective for the second image as well, the image state detection for the second and subsequent documents is performed. The search time for possible processing is shortened as compared with the case of the first original. Therefore, when inputting a plurality of pieces of image data, the processing time required for inputting is shortened.

As described above, it is possible to provide an image processing means for inputting and processing a plurality of originals at high speed.

[0056]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a flow chart showing the operation procedure of the image input method according to the present invention. FIG. 4 is a diagram showing a configuration example of an image processing apparatus for realizing the method according to the present invention.

First, the configuration of the image processing apparatus will be described with reference to FIG.

This apparatus comprises a scanner 400 and a memory 40.
1, control unit 402, encoding circuit 403, communication circuit 40
4, storage device 405, printer 406, frame memory 407, display 408, black run extraction circuit 700,
And a bus 409.

The black run extraction circuit 700 is means for detecting, when a predetermined number or more of black pixels are continuous, to assist tilt angle detection processing and the like, which will be described later. A detailed description will be given later with reference to FIGS. 6 and 7. Of course, the black run extraction circuit 700 is not an essential component in the present embodiment, and it goes without saying that the control unit 402 may perform the detection processing of the tilt angle.

The components of this apparatus will be described below.

The scanner 400 is a means for optically scanning the document surface and reading the image data.

The memory 401 is an area for storing image data read via the scanner 400 and used by various programs executed by the control unit 402. For example,
It is realized by an electronic device such as a RAM.

The control section 402 is means for performing various processing such as angle detection processing of image data stored in the memory 401 and rotation processing of image data. For example, CPU, R
It can be realized by an electronic device such as OM, RAM, various CMOSs, and a program stored in advance in the ROM.

The encoding circuit 403 is a means for encoding data in order to compress the input image data,
For example, it can be realized by electronic devices such as various CMOSs. The communication circuit 404 is means for transmitting and receiving encoded image data, and can be realized by electronic devices such as resistors, capacitors, and various CMOSs.

The storage device 405 is means for accumulating the encoded image data, and can be constituted by, for example, a magnetic disk, an optical disk or the like.

The printer 406 is a means for printing the image data on the memory, and a commercially available printer may be used.

The frame memory 407 is means for storing data corresponding to the display screen, and is realized by an electronic device such as a RAM.

The display 408 is means for displaying data on the frame memory, and is, for example, a CRT or EL.
It can be realized by a display or a liquid crystal display.

The bus 409 is a data communication means,
For example, realized by an electric harness. Next, the processing procedure of the image input method according to the present invention will be described with reference to the flowchart of FIG. 1 and FIG.

In this embodiment, a case will be described in which the present invention is applied to an electronic filing function for storing input image data in a storage device. Of course, the application example is not limited to this.

First, in step 100, the scanner 40
The image data read via 0 is stored in the memory 401. Here, it is assumed that the read image data is inclined with respect to a certain reference line.

Next, in step 101, the control unit 402
However, the inclination angle detection processing for the “horizontal writing character” is performed on the image data on the memory 401.

A method of detecting the tilt angle with respect to this horizontally written character will be described later.

When the tilt angle is detected, the read image is determined to be a horizontally written character.

Then, in step 102, the control unit 402 performs rotation processing on the image data by the detected tilt angle. The image data after the rotation processing is encoded by the encoding circuit 403, and in step 103, the data after the encoding processing is stored in the storage device.

Next, if there is a next image, step 1
At 00, the next image is input and a similar image inclination correction process is performed.

Next, in the case where the input document is "vertical writing characters", the angle detection process for horizontal writing characters is performed in step 101 on the image data input in step 100.

However, the input image data is
Since it is a vertically written character, the inclination angle cannot be detected (“impossible” in step 102), and the process branches to step 105. Here, since the angle detection process for the vertically written characters has not been performed on the image, the process proceeds to step 106.

At step 106, control unit 402 performs a tilt angle detection process for vertically written characters. The tilt angle detection processing for vertically written characters will be described later.

Since the target image is vertically written characters, the tilt angle is detected.

Next, in step 107, the control unit 4
The image data is rotated by the tilt angle detected by 02. For the image data after such rotation processing,
Encoding processing is performed by the encoding circuit 403, and step 1
At 08, the encoded data is stored in the storage device.

Next, when the next image exists, the process branches to step 110 and the data of the next image is input through the scanner 400. Then, in step 106, the inclination angle detection process for vertically written characters is performed on the input next image data. As shown above, when images of vertically written characters are input continuously,
For the first first image, processing for horizontal writing characters and processing for vertical writing characters are performed.
Inclination angle detection processing for vertically written characters is preferentially performed on the first and subsequent images.

As a result, when the same type of image is continuously input, the inclination angle can be detected for the second and subsequent images by a single process.

Therefore, since the time required for the image state detection processing for the second and subsequent images is shortened, the time required for total image input can be greatly reduced.

If a horizontally written character image is input during these processes, the inclination angle detection process for the vertically written characters becomes impossible in step 106, and the process branches to step 111. Thereafter, the above-described horizontally written characters are processed. The tilt angle detection process is performed. If the target image has not been subjected to the processing for horizontal writing (information on whether or not such processing has already been performed may be stored in, for example, the storage area of the control unit), step 101 is performed. Branch and perform horizontal writing angle detection processing.

When the process for the horizontally written characters is performed, the process branches to step 112 to give a warning to the operator that the tilt angle cannot be detected, and the series of processes is terminated. The warning is given by an alarm such as a chime, LE
It is conceivable that the display is such as D.

Although the present embodiment has been described with respect to the electronic file device, the printer 40 is replaced by the image registration process in step 103 and step 108 instead of the image print process.
It is easy to realize the function of the copying machine by printing the image data after the correction processing via 6, and similarly, the image registration processing in step 103 and step 108 is replaced with the image communication processing, and the correction processing is performed. It is easy to realize the function of the facsimile device by performing the encoding process on the subsequent image data by using the encoding circuit 403 and transmitting the encoded image data by the communication circuit 404. Is. Of course, the application example according to the present invention is not limited to this.

The image data after the correction processing is stored in the frame memory 407, and the display 408 is further stored.
It is also possible to realize a configuration in which the image after the correction processing is displayed on the display, and after the operator is asked to confirm, the storage, the printing, the communication and the like are performed.

Further, in the present embodiment, the scanner 400
It is also possible to further reduce the burden on the operator by automatically and continuously feeding the input original with a configuration having an automatic paper feeding function.

The following configuration is also possible.

That is, in the above embodiment, the back surface of the document pressing cover of the scanner 400 is painted black.

Next, at the time of image input, the operator inputs an image without designating the image size. Next, the scanner 40
The control unit 402 sequentially reads the image data of the entire screen loaded into the memory 401 via 0 from the image end.

A region of continuous black pixels (a pixel having a density equal to or higher than a certain threshold is called a "black pixel") is a region of continuous white pixels (a pixel which is not a black pixel). By detecting the coordinates of the changing point, specifying the image start position, and further detecting the coordinates of the point at which continuous white pixels change to the continuous black pixel region, by specifying the image end position, for example, JP-A-1-1065
Using the means shown in the prior art described in Japanese Patent No. 70,
It is possible to automatically detect the size of the input image and the position of the image on the platen.

By automatically detecting the size of the input image and the position of the image on the platen together with the inclination detection process, the image input apparatus with less burden on the operator can be obtained. Realization is possible.

The details of the image tilt angle detection processing may be the same as the method described in Japanese Patent Laid-Open No. 62-14277, for example. An embodiment of the process for detecting the inclination angle of a written character for an image will be described.

First, a method for detecting the inclination of an image of a horizontally written character will be described.

It is assumed that the image data read via the scanner 400 exists in the memory 401.

The control unit 402 sequentially reads this image data, and the contour points P1 and P at the lower right of the character in FIG.
The coordinates of 2, P3, etc. are detected.

The coordinates of the contour point P are (x, y). The processing for detecting the contour points will be described later.

Normally, in horizontal writing, the contour points on the same line are on the same straight line, as shown in FIG. Therefore, the inclination angle of this straight line becomes the inclination angle of the image.

If each contour point is projected in the same direction as this straight line, the number of projection points is reduced. That is, when the contour points are projected for each angle, it can be determined that the projection angle θ having the smallest number of projection points is the tilt angle of the image.

Further, the angle detection processing will be described in detail.

First, the projection process will be described.

The coordinates (x, y) of each contour point are projected in a direction inclined by θ degrees from the horizontal direction, and a coordinate point A intersecting with the left end portion of the entire image where X = 0 is obtained. When the y coordinate of this coordinate is AY, AY is calculated by the following formula.

AY = y−x · tan θ (Equation 1) Here, for example, the contour points P1, P2, and P3 are tilt angles θ.
If they exist on the same straight line, the contour points P1 and P
Projection points A1 and A of P2 and P3 with respect to a certain angle θ direction
2, A3 match.

The smaller the number of contour point projection points P in the angle θ direction, the greater the number of contour points of each character on a straight line inclined by θ degrees from the horizontal direction.

Let Nθ be the number of projection points A in a certain angle θ direction. By changing θ within a certain range, Nθ for each angle is stored, the value of Nθ having the largest number is set as NMAX, and the value of Nθ having the smallest number is set as NMIN.

The larger the difference between NMAX and NMIN, the more N
It is particularly likely that a contour point exists on a straight line having an inclination angle of the angle at which MIN is detected.

Therefore, the criterion for determining whether or not the angle can be detected is to calculate the value NMAX / NMIN, and make it correspond to whether or not the calculated value exceeds a predetermined threshold value. Normally, when this threshold value satisfies the criterion of about “2”, the angle at which NMIN is detected is set as the tilt angle of the image.

With the above method, it is possible to detect the tilt angle for horizontally written characters.

A case where the above-described inclination angle detection method for images of horizontally written characters is applied to an image of vertically written characters as shown in FIG. 3 will be described below.

In the case of a vertically written character, as shown in FIG. 3, since the size of the character is different, the contour point Q of the character on the adjacent line is slightly deviated.

Therefore, no matter how many horizontal points are projected, many contour points Q do not lie on the same straight line.

Therefore, the difference between NMAX and NMIN becomes small, and the angle cannot be detected. For the above reasons, the tilt angle detection method for a horizontally written image cannot be applied to a vertically written character image.

Next, the image inclination angle detection processing for vertically written characters will be described.

First, similarly to the inclination angle detection processing for an image of a horizontally written character, the control unit 402 sequentially reads the image data in the memory 401 read by the scanner 400, and the contour point at the lower right corner of the character is read. Detects Q1, Q2 and Q3.

The coordinates of this contour point are defined as (x, y). Since the processing for detecting the coordinates of the contour points is similar to the inclination detection processing for horizontal writing characters, if the inclination detection processing for horizontal writing has already been performed, the result can be diverted and can be used again. Processing is unnecessary.

Normally, in the case of vertically written characters, as shown in FIG. 3, the contour points of the characters on the same line exist on a certain straight line. Therefore, the inclination angle of this straight line becomes the inclination angle of the image.

If each contour point is projected in the same direction as this straight line, the projection points of the contour points overlap each other, and the number of projection points decreases. That is, when the contour points are projected for each angle, it can be determined that the projection angle θ having the smallest number of projection points is the tilt angle of the image.

Next, the projection process will be described.

The coordinates of each contour point are projected in a direction inclined by θ degrees from the vertical direction, and a coordinate point B intersecting with the upper end of the whole image where y = 0 is obtained. When the x coordinate of this coordinate is BX, BX is obtained by the following equation.

BX = x−y · tan θ (Equation 2) Then, similar to the angle detection process for the horizontally written characters, the projection process is performed for each angle, and the number of projection points B for each angle is counted.

Next, the maximum value of the projection points is NMAX and the minimum value of the projection points is NMIN.

Further, a value NMAX / NMIN is obtained, and if this value is equal to or larger than a predetermined threshold value,
It is determined that the angle detection is performed, and the angle at which NMIN is detected is set as the tilt angle of the image.

Incidentally, the threshold value is usually about “2” as in the case of the inclination detection process for horizontally written characters. With the above processing, it is possible to detect the tilt angle with respect to the vertically written characters.

Next, a case in which the image inclination detection method for vertically written characters is applied to an image for horizontally written characters will be described below.

As shown in FIG. 2, since horizontal writing characters have different sizes for each character on each line, vertical writing
The contour points of the characters are not aligned.

When the tilt angle detection method for vertically written characters is applied to an image of horizontally written characters, many contour points Q do not lie on the same straight line no matter what angle the image is projected. Therefore, the difference between NMAX and NMIN becomes small, and the angle cannot be detected.

For the above reason, the angle detection process for vertically written characters cannot be applied to the image of horizontally written characters.

Next, how to obtain the contour points P and Q in FIGS. 2 and 3 will be described.

FIG. 5 shows the image data read in the memory 401. The shaded area indicates an area that is a black pixel when the image is read through the scanner 400, and the other area (illustrated by “×”) is a white pixel area when the image is read through the scanner 400. .

Next, the procedure for detecting the contour points P and Q will be described.

First, the control unit 402 determines that one of the entire images
The image data of the highest line is read and copied to the buffer area in the memory 401.

When the copying of the data of one line is completed, the control unit 402 sequentially reads the data of the next line of the input image from the memory 401 from the left pixel.

At this time, the data of the previous line copied from the buffer area of the memory 401 is read at the same time. When the pixel on the left side of the processing target pixel is a black pixel, the upper pixel of the processing target pixel read from the buffer area is a black pixel, and the processing target pixel is a white pixel, the control unit 402 sets the processing target pixel to a black pixel. Convert. In other cases, the color conversion is not performed and the process for the next pixel is performed, and the same color conversion process as described above is performed.

Here, even when the pixel on the left side of the pixel to be processed is converted into a black pixel in the pre-processing, the pixel to be processed is converted to a black pixel in the same manner as when the pixel on the left side is originally a black pixel. . Then, the data subjected to the color conversion processing is written in the buffer area of the memory 401.

Here, it is not necessary to use the data written in the buffer area in the previous processing and which is already read for processing the current target line.

Therefore, the color-converted data can be written in this processed area to effectively use the buffer area.

The above process is repeated to complete the process for all lines.

Explaining with reference to FIG. 5, the input image data is only copied to the buffer area up to the 4th line, but the pixel of a on the 5th line is the pixel on the left side and the pixel on the upper side. Since it is a black pixel, it is converted into a black pixel.

The pixel b is also converted into a black pixel because the pixel on the left side and the pixel on the upper side are black pixels. By repeating the above processing, all the pixels indicated by "x" in FIG. 5 are converted into black pixels.

When the pixels to be processed are continuously white pixels and the pixels on the line above the processed pixels read from the buffer area are continuously black pixels, the number of these consecutive pixels is counted.

At this time, if a black pixel appears in the pixel to be processed, the counting process is stopped and the color conversion process is started.

When the pixel on the previous line reaches the position where the pixel is a white pixel and the number of consecutive pixels is within a preset range, the position of the last pixel of the consecutive lines is set to the outline of the character. The lower right point of the component is the contour point. Figure 5
In, P is that position.

The control unit 402 writes the x and y coordinates of this point in the work area of the memory 401 and uses them in the projection process for detecting the tilt angle later. The above processing is repeated for one image.

With the above processing, it is possible to detect the contour point at the lower right of the character.

Further, this continuous pixel number L is defined as the "contour size" of the character, and when the first image is input, not only the coordinate position of the contour point but also the character size L is written in the work area and controlled later. It is also possible to realize an image input method in which the unit 402 checks the distribution range of the character size L and determines the detected character range in consideration of the distribution range when the second and subsequent images are input.
In addition, the above-mentioned embodiment or JP-A-62-1427.
In the method described in Japanese Patent Publication No. 7, the image tilt detection processing is realized by software processing by a general processor, but a part of the processing may be implemented as dedicated hardware to perform the processing. Needless to say.

As another image inclination angle detection processing method, the control unit 402 detects the coordinates of a point where a predetermined number of black pixels in the image are continuous, and the coordinate point is set in the horizontal direction shown in FIG. It is also possible to detect the inclination angle of the horizontal line in the image by treating it in the same manner as the contour point of the character inclination detection processing.

This method is effective for an image in which there are many horizontal lines in the image, for example, a table or a line figure exists.
Therefore, in addition to the vertical writing character and the horizontal writing character, a tilt angle detection processing method using continuous black pixels is prepared as the tilt detection processing for the horizontal line. It is also possible to extend the method of changing the priority order among the three methods and perform the tilt angle detection while changing the priority order among the three methods.

The line segment inclination detection process will be described in more detail below with reference to the drawings. Here, a case where continuous black pixels are referred to as “black run” and the black run is extracted from the input image by the dedicated hardware will be described.

FIG. 6 shows an example of black run component extraction processing.
FIG. 7 shows an example of the configuration of the black run extraction circuit.

In FIG. 7, reference numeral 700 is a black run extraction circuit, which is connected to the bus 409 of FIG.

The main black run extraction circuit 700 includes a register 70.
1, parallel-serial conversion circuit 702, counter 70
4, OR circuit 703, counter 705, control circuit 70
7 and a buffer 708. Also,
A bus 409 is connected to the black run extraction circuit 700.

The components will be described below.

The register 701 is means for setting the number of continuous black pixels to be extracted, and can be realized by various CMOSs or the like.

The parallel-serial conversion circuit 702 is means for converting the image data read via the bus 409 in parallel in units of words in units of pixels, and can be realized by various CMOSs, for example.

The OR circuit 703 is means for performing a logical sum operation process, and can be realized by various CMOSs, for example.

The counter 704 is a means for counting the number of pixels in the horizontal direction during processing, and can be realized by various CMOSs, for example.

The counter 705 is a means for counting the number of continuous black pixels, and can be realized by various CMOSs, for example.

The comparator 706 is means for comparing the output value of the register 701 with the output value of the counter 705, and can be realized by various CMOSs or the like.

The control circuit 707 is means for controlling the operation of the entire black run extraction circuit, and includes, for example, a CPU, a ROM in which a program created in advance is stored, a RAM,
It is realized by various CMOSs.

The buffer 708 is means for temporarily storing data to be output to the bus 409, and is realized by, for example, a RAM or the like.

Next, the extraction process of the black run component will be described.

It is assumed that the memory 401 stores image data of an inclined line segment (a line segment having a step) read by the scanner 400 as shown in FIG.

First, the control unit 402 sets the number of consecutive black pixels to be detected in the register 701 through the bus 409 before starting the black run extraction processing operation. This value
It will be referred to as "LrMin".

Next, the black run extraction circuit 700 operates in the memory 4
The image data in 01 is sequentially read in word units from the upper left to the lower right through the bus 409 to the parallel / serial conversion circuit 702. Here, the parallel-serial conversion circuit has a function of performing so-called parallel-serial conversion in accordance with a clock supplied from the control circuit 707 and outputting pixel-unit data.

At this time, the counter 704 has the control circuit 70
A count-up is performed in accordance with a clock supplied from 7 and the number of pixels in the horizontal direction to which the pixel currently being processed corresponds is stored.

Then, the parallel-serial conversion circuit 705
If the pixel output by is a black pixel, the counter 705
Counting up is performed, and in the case of a white pixel, the OR circuit 70
By 3, the counter 705 is reset.

Therefore, the count-up is continued only when the black pixels are continuous. When black pixels are continuous and the value of the counter 705 becomes equal to LrMin preset in the register 701,
The comparator 706 outputs a signal to the control circuit 707.

When the control circuit 707 receives the signal, the control circuit 707 transfers the value of the counter 704 indicating the horizontal position of the pixel currently being processed to the buffer 708.

This horizontal position is R in FIG.
The x-coordinate of each point of 1, R2, and R3 is shown. Then, the value of this horizontal position is stored in the memory 4 via the bus 409.
Write to the work area 01.

At this time, the control unit 402 also writes the number of vertical lines currently being processed in the work area in the memory 401. The number of lines in the vertical direction currently being processed indicates the y coordinate of R1, R2, and R3 in FIG.

Image 1 taken from the scanner 400
The above-mentioned processing is performed after the data for one sheet is input to the black run extraction circuit. When this process ends,
In the work area in the memory 401, as shown in FIG.
The x and y coordinates of points R1, R2 and R3 where black pixels are continuous for LrMin or more are written.

Upon completion of the above processing, the control unit 402
Performs projection processing for tilt angle detection.

The coordinates of each detection point written in the work area are projected in the direction inclined by θ degrees in the horizontal direction as in the case of the horizontal writing process, and the coordinates of the intersection C with the y axis are obtained.

When the y coordinate of this point C is CY, CY is given by the following equation.

CY = y−x · tan θ (Equation 3) At a certain projection angle θ, the control unit 402 reads the coordinates of each detection point R1, R2, R3 from the work area in the memory 401, and the projection point C Calculate the coordinates.

When the calculation of the coordinates of the projection points of one detection point is completed, the number Nθ of projection points at a certain projection angle θ is counted.

Thereafter, similar processing is performed by changing the projection angle θ, and Nθ for each angle is stored. Furthermore, in Nθ for each angle, the maximum value is NMAX and the minimum value is NMIN.
And If NMAX / NMIN exceeds a predetermined threshold, record the angle at which NMIN was detected. The threshold value is preferably about "2".

The angle at which this NMIN is detected represents the inclination angle of the straight line when the detection points R1, R2 and R3 in FIG. 6 are on the same straight line.

This is because there are many horizontal lines in an image having a table, a drawing, etc., which is equal to the inclination of the horizontal line in the original image captured by the scanner 400, and therefore, there are many horizontal lines in these images. By examining the tilt angle, the tilt angle of the image can be detected.

In the above inclination angle detection processing, the black run component extraction processing having the image processing function is performed by dedicated hardware, whereby the processing speed can be further improved.

The extraction of the coordinates of the contour points of the character in FIG. 5 can be speeded up by the same hardware.

When the first image is input, the control unit 40
2 may be a processing method in which the thickness distribution of the line segment existing in the first image in the memory 401 is investigated and used.

In this case, it is preferable that only the first image is processed in such a manner that the correct thickness of the horizontal line in the image is investigated by inputting it in the standard state without tilting.

Further, when the operator inputs a certain set of images at a time into the image input device that automatically corrects the input image, the direction in which the characters are written, the size of the characters, the line segment, and the like. It is also possible to adopt a method in which a part of the characteristic information of the image including the thickness of the image is input and used in combination with the above processing method.

Further, the characteristic of the name of a specific image may be stored in advance and the operator may select it when inputting the image. For example, when an image of a newspaper is input, the character size to be detected, etc., corresponding to the newspaper is registered, and when the operator sets that the input image of the newspaper is set, the operator can respond in advance to this. That is, the configuration is such that predetermined parameters are selected.

As described above, the present invention has an important effect that the input operation becomes easy and quick when performing the image input operation. Further, the present invention is also highly applicable to various information processing apparatuses including facsimile machines and copying machines.

[0190]

According to the present invention, when a plurality of images are input to various information processing apparatuses, if the same type of images are continuously input, it takes time to extract the features of the second and subsequent images. Since it is reduced, the time required for input is greatly reduced.

[Brief description of drawings]

FIG. 1 is a flowchart showing a processing procedure of an image input method according to the present invention.

FIG. 2 is an explanatory diagram of a tilt angle detection process for horizontally written characters.

FIG. 3 is an explanatory diagram of a tilt angle detection process for vertically written characters.

FIG. 4 is a diagram showing a configuration example of an image input device according to the present invention.

FIG. 5 is an explanatory diagram of an example of character contour point detection processing according to the present invention.

FIG. 6 is an explanatory diagram of a black run extraction process.

FIG. 7 is an example of a configuration diagram of a black run extraction circuit.

[Explanation of symbols]

400 ... Scanner, 401 ... Memory, 402 ... Control unit,
403 ... Encoding circuit, 404 ... Communication circuit, 405 ... Storage device, 406 ... Printer, 407 ... Frame memory, 4
08 ... Display, 409 ... Bus, 700 ... Black run extraction circuit, 701 ... Register, 702 ... Parallel-serial conversion circuit, 703 ... OR circuit, 704 ... Counter, 7
05 ... Counter, 706 ... Comparator, 707 ... Control circuit,
708 ... buffer

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Masaaki Fujinawa 2880, Kozu, Odawara-shi, Kanagawa Stock Company Hitachi Storage Systems Division

Claims (5)

[Claims] 1. An image input method for scanning image data by optically scanning a paper surface, wherein at least one of image data extracted at a previous image input process when inputting and reading a plurality of originals. An image input method characterized in that the image input processing of this time is performed using at least one of the characteristic information of the above. 2. The information according to claim 1, wherein the characteristic information is information including a direction in which a character is written on a document and a size of the character, and the tilt angle of the document is detected based on the information. An image input method characterized in that a rotation process corresponding to the angle is performed to correct a tilt angle of an image input this time. 3. An image inputting method for optically scanning a paper surface to read image data, wherein a plurality of originals having the same at least one of information including a character direction, a character size, and a line segment thickness is used. When inputting one image, the process used in the image input process for the first image is omitted in the image input process for the second and subsequent images, and compared with the time required for the image input process for the first image. An image input method characterized in that the time required for the image input processing for the second and subsequent images is shortened. 4. An image input method for scanning image data by optically scanning a paper surface, when inputting a plurality of originals on which an image having a certain directionality is drawn, The inclination angle of the original is detected based on the directionality extracted at the time of the image input processing, and the rotation angle corresponding to the angle is detected to correct the inclination angle of the image input this time. Image input method. 5. Image processing having an image input means for taking in image data, a storage means for storing the image data, and a function for extracting image characteristic information including directionality from the stored image data. And a correction processing unit having a function of detecting a tilt angle from a certain reference line, performing rotation processing, and performing tilt angle correction processing based on the image feature information. Input device.