JPH0787315A - Image resolution adaptive controller - Google Patents

Abstract

PURPOSE:To shorten the communicating time and to reduce the capacity of a storage part by grasping the local complicatedness for quantization in order to decide the accurate image resolution and to reduce the code quantity occurring as a low resolving level in a low information density state. CONSTITUTION:A changing point detecting part 710 detects a point of time when the pixels of the main scanning direction binary data 301 are changed from white to black or vice versa. A specific run length extracting part 720 calculates the number of continuous white or black pixels at the changing point of time detected by the part 710 and decides each run length. Thus a set specific run length and an array pattern are extracted from those run lengths. A specific run length counter 730 counts those specific run lengths to obtain a specific run length pattern number 731 that shows the information density, i.e., the quantitative value of the local complicatedness of an image. This quantitative value is compared with the set reference value at an image resolution deciding part 740. Thus the image resolution is decided for a read original 10.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a resolution adaptive control apparatus, and more particularly to a facsimile apparatus for binarizing and storing / transmitting an original image including at least one of a character, a line figure and a background, and a photograph as a halftone. Alternatively, the present invention relates to a resolution adaptive control device that optimizes the resolution, density, and filtering processing in halftone processing in a digital image reading apparatus.

[0002]

2. Description of the Related Art Generally, an image of an original image of this type is composed of pixels arranged in a grid pattern in the X and Y directions. For example, in a facsimile machine, the resolution is defined by the pixel density (number of pixels / mm) in the main scanning direction (X direction) and the scanning line density in the sub scanning direction (Y direction).
The density used together with the resolution is a measure of lightness and darkness depending on the degree of inclusion of the white level and the black level even if the resolution is the same, and the difference in the filtering conditions of the filtering process is in the reproduction of a halftone original image such as a photograph. Set the size of the emphasis on blurring, characters, and line shapes.

Regarding the operation of the conventional facsimile machine,
This will be described with reference to FIG.

As shown in FIG. 3, in the conventional facsimile apparatus, the document reading section 20 reads the document 10 from the multi-valued data 201.
, And the image is processed by the binarized image processing unit 30 to be converted into the black and white binary data 301, and then the data compression unit 40 performs data compression by the coding process to obtain the compressed data 401. These data are stored in the storage unit 50 or sent to the line by the transmission unit 60. In this case, in the data compression unit 40, MH (Modified
Huffman) and MR (Modified Read) are used to compress data to ensure efficient storage and transmission. Further, the resolution of the binary data 301 in the main scanning direction and the sub scanning direction is automatically changed as necessary to reduce the image data.

Here, some of conventional conventional methods for compressing data by controlling resolution while paying attention to image redundancy will be described.

For example, in the "information density adaptive resolution conversion method" of Japanese Patent Laid-Open No. 59-11068, the number of change points from white to black and from black to white of binary data in the scanning direction is counted, and this change is counted. Data is compressed by comparing the score with a predetermined reference value and changing the resolution according to the determination result.

Further, in the "Automatic linear density switching control system in a one-dimensional encoding facsimile apparatus" of Japanese Patent Laid-Open No. 58-54777, one-dimensional encoding processing (MH system encoding) is performed on binary data in the scanning direction. Processing), and the resolution is controlled based on the determination result obtained by comparing the code amount with a predetermined reference code amount.

Further, in the "Automatic linear density switching control system in a two-dimensional encoding type facsimile apparatus" of Japanese Patent Laid-Open No. 58-54778, two-dimensional encoding processing is performed on binary data in the main scanning direction and the sub-scanning direction. Then, the resolution is controlled based on the determination result obtained by (MR system encoding process) and comparing the code amount with a predetermined reference code amount.

[0009]

Each of the above-mentioned conventional resolution control methods has the following drawbacks.

The resolution control system based on the number of change points per line according to the "information density adaptive resolution conversion system" disclosed in Japanese Patent Laid-Open No. 59-11068 has a drawback that the local complexity of an image cannot be inferred.

FIG. 4 is an explanatory diagram of the number of change points in the case of not being locally complicated (a) and in the case of locally complicated (b). 4 (a) and 4 (b), the number of change points is 10. However, (b) is locally more complicated and has a higher information density. In other words, the local complexity cannot be inferred only by the number of change points.

Further, in the "Automatic linear density switching control system in a one-dimensional encoding type facsimile apparatus" of Japanese Patent Laid-Open No. 58-54777, a resolution control system based on a one-dimensional encoding process (MH system encoding process) is used. Since the coding process itself is devised so that the code amount does not increase so much, there is a drawback that the information density and the code amount are not exactly proportional. FIG. 5 shows a part of the termination code of the one-dimensional coding method. As can be seen from FIG. 5, especially when the black run length is 2 to 3, it is a 2-bit code word. Therefore, even if there are two or three white or black pixels, the total code amount can be suppressed.

Further, even in the resolution control system of "Automatic linear density switching system in a two-dimensional encoding facsimile apparatus" of Japanese Patent Laid-Open No. 58-54778, the same thing as in the case of the above-mentioned one-dimensional encoding process is performed. However, there is a drawback that the information density and the code amount are not exactly proportional. In the two-dimensional coding method, the first one line that serves as a reference is one-dimensionally coded, and therefore the same as the above can be said. Further, since the subsequent lines are encoded by changing positions with reference to the previous lines, respectively, a high correlation between the information density in the main scanning direction and the code amount cannot be ensured.

An object of the present invention is to eliminate the above-mentioned drawbacks, to accurately grasp and quantify local complexity, and to perform resolution control capable of resolution control while ensuring a high correlation between information density and code amount. To provide.

[0015]

SUMMARY OF THE INVENTION The device of the present invention comprises a character,
In a facsimile device capable of binarizing an original image including at least one of a line figure, a background and a photograph into a binary image and variably controlling the resolution in at least one of a main scanning direction and a sub scanning direction, The binary image of the original image is quantified based on the number of patterns having a specific run length included in the binary image obtained by scanning the original image in the main scanning direction. It has a configuration in which the resolution in at least one of the scanning direction and the sub-scanning direction is adaptively controlled sequentially for each scan or in page units.

Further, the apparatus of the present invention is a facsimile apparatus which binarizes an original image including at least one of a character, a line figure, a background and a photograph to form a binary image and stores and transmits the binary image. Alternatively, in a digital image reading apparatus, the number of patterns having a specific run length included in the binary image obtained by scanning the original image in the main direction is counted, and the binary image is locally quantified based on the counted value. In accordance with complexity, the resolution and density in the reproduction of the characters, line graphics and background of the original image are sequentially adaptively controlled for each main-direction scan, and the blurring or emphasis in the halftone reproduction of the photograph is targeted. The configuration is such that the filtering process is sequentially adaptively controlled for each main-direction scan or for each page.

Further, the apparatus of the present invention is a facsimile apparatus for performing the image processing of the original image, in which the number of patterns having the specific run length is counted and the binary image is locally quantified based on the counted value. By comparing the complexity with a preset reference pattern, it is determined whether the read original image is a region including characters and line graphics, a region including a photograph, or a region including a background. On the basis of the above, there is a configuration in which the size of the resolution in reproducing the original image and the density expressing light and dark and the averaging condition in the averaging process are sequentially adaptively controlled for each scan or in page units.

[0018]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, the present invention will be described with reference to the drawings. FIG. 1 is a configuration diagram of a first embodiment of the present invention. In the first embodiment, an original reading unit 20 that reads the original 10 as a grayscale image and binary data 30 that reads the read grayscale image.
A binarized image processing unit 30 that performs conversion processing to 1, a resolution automatic determination circuit 70 directly related to the present invention, and a resolution switching image process that sequentially adaptively controls the resolution for each main scanning according to the determination result from the resolution automatic determination circuit 70. Unit 80, data compression unit 40 for compressing binary data 301, and storage unit 50 for storing binary data 301 or compressed data 401.
And a transmission unit 60 for transmitting the compressed data 401 of the binary data to the line.

Further, the automatic resolution determination circuit 70 includes a changing point detecting section 710 for detecting a changing point of binary data, a specific run length extracting section 720 for extracting a predetermined specific run length pattern, and an extracting step. The specific run-length counter 730 that counts the number of specified run-length patterns and outputs the result as the specific run-length pattern number 731, and the specific run-length pattern number 731 is determined in advance as the specific run-length pattern number 731. It has a resolution determination unit 740 that compares the reference value with the resolution and determines an appropriate resolution.

Consider, for example, a case where a facsimile machine scans an original image in the main scanning direction. Manuscript 10
Is read by the document reading unit 20 and output as multi-valued data 201. The binary image processing unit 30 converts the multivalued data 201 into binary data 301. The automatic resolution determination circuit 70 determines the information density of the binary data 301, so-called “local complexity of image”, and sends the determination result 701 to the resolution switching image processing unit 80. Resolution switching image processing unit 80
Is a resolution switching output 80 adjusted and converted to an appropriate resolution.
It is sent as 1. Furthermore, this resolution switching output 801
Data is compressed by the data compression unit 40 and stored in the storage unit 50 or sent to the transmission unit 60.

Next, the operation of the automatic resolution determination circuit 70 will be described.

The change point detector 710 searches the binary data 301 in the main scanning direction for a change point at which a pixel changes from white to black or from black to white. The specific run length extraction unit 720 calculates the continuous number of white or black pixels from the change points of the binary image data 301 to obtain the respective run lengths, and a specific preset length set from these run lengths is calculated. The run length and the array pattern are extracted, and the number of them is counted by the specific run length counter 730. The value thus obtained is the specific run length pattern number 731, and can be said to be a quantitative numerical value of the information density, that is, the local complexity of the image. Then, the resolution of the read original 10 is determined by comparing this numerical value with a predetermined reference value.

Characters, figures, and photographs written on a document are reflected in image data obtained by scanning them, and take values of black and white. If there are few line elements that cross the scanning line, the number of change points will also be small. In the case of small characters or Chinese characters with many strokes, the width of the line elements of the characters and the spacing between the line elements become small. For example, when scanning with a resolution of 8 pixels / mm, there are many run length patterns of several pixels or less. Become. Further, in a facsimile machine, pseudo halftone reproduction processing such as an area gradation method is used for processing halftone portions such as photographs.

As a method of this pseudo halftone reproduction processing,
A method called an area gradation method is usually used. This area gradation method expresses ON (black) in an image when it is desired to reproduce a grayscale image by using a display device or a printer device capable of displaying only binary values in which 0 is turned off and 1 is turned on. This is to reproduce the gradation by changing the ratio of 1.
The error diffusion method as a decision method is generally used for the systematic dither method and the condition as an independent decision method known as the density pattern method or the dither method.

In the density pattern method, one pixel of the original image is made to correspond to a sub-matrix of n dots in each of the vertical and horizontal directions, and the density of the pixel is reproduced by the black occupation area ratio included in the sub-matrix.

Further, in the systematic dither method, a sub-matrix called a dither matrix is superimposed on the entire original image, and compared with the threshold value given to the corresponding dither matrix, and binarization is performed according to the magnitude.

Further, as a conditional decision method of the dither method in a broad sense, there are an error diffusion method and an average error minimum method, but these methods can be said to be equivalent methods as a result except for handling the edge of the image. The error diffusion method convolves the error between the binarization result of the pixel in the sub-matrix, which is a weighting filter (matrix), and the density of the pixel so that the difference in density between the original image and the output image is reduced as an average. The density of the pixel of interest is compensated for and threshold processing is performed.

The above-mentioned pseudo halftone reproduction processing is described in detail, for example, in "Image Analysis Handbook" supervised by Takagi and Shimoda, The University of Tokyo Press, 1991 and the like.

In the halftone reproduction processing by the above-mentioned methods, in the case of the density pattern method, the appearance period of the white pixels and the black pixels depends on the unit of the sub-matrix. For example, when the size of the sub-matrix is 4 × 4 pixels. Indicates that there are many run length appearance patterns (patterns) of 4 pixels or less. Further, in the error diffusion method, the amount of error when binarizing a pixel in the vicinity of the pixel of interest is weighted to correct the density of the pixel of interest and the threshold value is binarized. The periodicity due to the coefficient matrix is not noticeable, but there is a pattern feature that a run-length pattern of several pixels or less appears according to the density of the original pixel, and the systematic dither method also has the original image of the dither matrix. There is a run length pattern characteristic due to the polymerization of.

Due to the above-mentioned characteristics concerning the appearance run length, it is about 1 to 4 pixels for a document having many small characters or complicated line figures requiring fine resolution, and a document having a halftone portion such as a photograph. The local information density, that is, the local complexity in the original can be estimated by extracting the specific pattern having the run length. Also, the optimum resolution can be estimated by counting the number of the specific patterns and comparing them with a preset reference value. Further, if the local complexity in the main scanning direction is taken into consideration even in the continuity in units of rows, it becomes stronger against noise and more accurate determination can be performed.

In the above description, the resolution switching image processing unit 80 adjusts the image to be stored / transmitted in scanning line units or page units to an appropriate resolution according to the resolution determination result 701 in one embodiment. However, there are various kinds of variations, and for example, the document reading unit 20 causes the document 1
It is obvious that the process of re-reading 0 at an appropriate resolution and the designation of the reading resolution of the next line or page can be easily performed.

FIG. 2 is a block diagram of the second embodiment of the present invention. In the second embodiment, the original reading unit 20 that reads the original 10, the binarized image processing unit 30 that converts the read multivalued data 201 of the grayscale image into a binary image, and the type of the read area are set. An area determination circuit 90 for determination, a density switching control 600 for controlling density switching based on the determination result 901 from the area determination circuit 90, an image processing method switching control section 610 for controlling image processing method switching, and resolution switching. A resolution switching control unit 620 for performing control, a resolution switching image processing unit 690 for performing resolution switching processing, a density conversion processing unit 630 for changing the density of the read grayscale image, and smoothing processing for the grayscale image. A smoothing processing unit 640, a sharpening processing unit 650 that emphasizes line elements of a grayscale image, a binarized image processing method unit (A) 660 that can perform binarization using different image processing methods, and Nekaga processing system unit (B) 670
A selection circuit 680, a data compression section 40 for compressing binary image data, a data storage section 50 for storing binary image data or compressed data, and a transmission for sending the image data to the line. And a part 60.

The area discrimination circuit 90 includes a binary image data change point detection unit 910, a specific run length extraction unit 920 for extracting a predetermined specific run length pattern, and a specific run length pattern. A specific run length counter 930 that counts the number of the specific run length patterns, and an area determination unit 940 that classifies the type of the target area by comparing the specific run length pattern number 931 with a predetermined reference value.

Similar to the first embodiment shown in FIG. 1, consider the case where a facsimile machine processes image data in the main scanning direction.

The original 10 is read by the original reading unit 20,
The binarized image processing unit 30 outputs a binarized image 301 that has not been binarized for region determination. Area determination circuit 9
0 is used to estimate the type of document in the target area, and the determination result 901 is notified to the density switching control unit 600, the image processing method switching control unit 610, and the resolution switching control unit 620.

The density conversion processing unit 630 converts the density of the multivalued data 201 as a grayscale image read by the document reading unit 20 according to an instruction from the density switching control unit 600.
The smoothing processing unit 640 changes the parameter for smoothing the grayscale image according to an instruction from the image processing method switching control unit 610. Further, the sharpening processing unit 650 changes the parameter for sharpening the grayscale image. Further, the binarized image processing method unit 660 and the binarized image processing method unit (B) 670 respectively select appropriate ones from those binarized by different image processing methods by the selection circuit 980. The two binarized image processing methods described above are set in advance based on the operation purpose. The resolution switching processing unit 690 includes a resolution switching control unit 6
According to the instruction of 20, the resolution of the binary image can be converted in at least one of the main scanning direction and the sub scanning direction. Further, this image data is compressed by the data compression unit 40 and either stored in the data storage unit 50 or sent to the line by the transmission unit 60.

Next, the operation of the area discrimination circuit 90 will be described. The change point detection unit 910 searches for a change point in the binary data 301 where the pixel changes from white to black or from black to white in the main scanning direction. The specific run length extraction unit 920 calculates the number of consecutive white or black pixels from the change point of the binary data 301, and extracts a specific run length or array pattern set in advance, A counter 930 counts those numbers. As described in the first embodiment, the extracted specific run lengths are often present in fine characters and halftone portions. Therefore, depending on the setting method, it is possible to determine whether there are fine characters in the target area, there are halftones such as photographs, or whether the area corresponds to the background.

Next, what kind of processing is to be performed according to the determined area will be described with an actual example. For example, in an area with many small characters, conversion processing is performed to increase the density of the read grayscale image, smoothing processing is suppressed, strong parameters are used for sharpening processing, and binarization processing is performed using moving average. The binarized image processing method unit (A) 660 for performing threshold processing by the method is selected. Furthermore, regarding the resolution,
Use fine resolution.

In a large character area or background area, the absolute amount of data is reduced by simple thinning or thinning by averaging processing in at least one of the main scanning direction and the sub scanning direction in terms of resolution. Is performed by the resolution switching image processing unit 690.

Further, a binarized image processing method (B) in which density conversion processing is not performed in the photograph portion, strong parameters are used in smoothing processing, sharpening processing is suppressed, and halftone reproduction processing by an error diffusion method is performed. 670 is selected and the resolution is set finely.

In addition, when characters are printed on a light-colored original, it may be possible to reduce the density of the background portion to secure the contrast with the character portion.

In this way, the area determination circuit 90 can infer what kind of original the target area or its page is, and select the image processing suitable for it.

[0043]

As described above, according to the present invention, in a facsimile apparatus or a digital image reading apparatus which reads an original image and outputs it as binary data, the size and amount of characters or figures in an original to be read, or halftone of a photograph or the like. The amount of parts is quantified as local complexity, and by comparing this with a predetermined reference value, the necessary and sufficient resolution can be determined extremely accurately, and the code amount that occurs as low resolution when the information density is small. Therefore, the communication time can be shortened and the capacity of the storage unit can be reduced.

[Brief description of drawings]

FIG. 1 is a configuration diagram of a first embodiment of the present invention.

FIG. 2 is a configuration diagram of a second embodiment of the present invention.

FIG. 3 is a block diagram of a conventional facsimile apparatus.

FIG. 4 is an explanatory diagram of the number of change points in the case of not being locally complicated (a) and the case of being locally complicated (b).

FIG. 5 is a diagram showing a part of a termination code of a one-dimensional encoding system.

[Explanation of symbols]

 10 document 20 document reading unit 30 binarization image processing unit 40 data compression unit 50 data storage unit 60 transmission unit 70 automatic resolution determination circuit 80 resolution switching image processing unit 90 area determination circuit 600 density switching control unit 610 image processing method switching Control unit 620 Resolution switching control unit 630 Density conversion processing unit 640 Smoothing processing unit 650 Sharpening processing unit 660 Binary image processing method unit (A) 690 Binary image processing method unit (B) 680 Selection circuit 690 Resolution switching Image processing unit 710,910 Change point detection unit 720,920 Specific run length extraction unit 730,930 Specific run length counter 740,940 Resolution determination unit

Front page continuation (72) Inventor Daisuke Yagi 4-2 Shimamata, Kakegawa, Shizuoka Prefecture Shizuoka NEC Corporation

Claims (3)

[Claims] 1. An original image including at least one of a character, a line figure, a background and a photograph is binarized into a binary image, and the resolution of at least one of a main scanning direction and a sub scanning direction is variably controlled. In a possible facsimile apparatus, the number of patterns having a specific run length included in the binary image obtained by scanning the original image in the main direction is counted, and the binary image of the original image is quantified based on the counted value. Corresponding to this, the resolution adaptive control device is characterized in that the resolution in at least one of the main scanning direction and the sub-scanning direction is adaptively controlled for each scanning or for each page. 2. A facsimile device or a digital image reading device for binarizing an original image including at least one of characters, line graphics, background and photograph to form a binary image and storing and transmitting the binary image. In, corresponding to the local complexity of the binary image obtained by counting the number of patterns having a specific run length included in the binary image obtained in the main scanning of the original image and quantified based on the counted value. And adaptively control the resolution and density in the reproduction of the characters, line figures, and background of the original image for each main-direction scan, and the main direction is the filtering process for the blurring or enhancement in the halftone reproduction for the photograph. A resolution adaptive control device characterized by performing adaptive control for each scan or page by page. 3. A facsimile apparatus which performs image processing of the original image, presets a local complexity of the binary image, which is obtained by counting the number of patterns having the specific run length and quantifying based on the counted value. The original image read based on the determination result after determining whether the read original image is a region including characters and line graphics, a region including a photograph, or a region including a background by comparing with the reference pattern 3. The resolution adaptive control apparatus according to claim 2, wherein adaptive control is sequentially performed for each scan or for each page for the density representing the magnitude of the resolution and the density expressing the darkness and the filtering condition in the filtering processing in the reproduction of the above.