JP4177417B2 - Dropout color identification processing program, method and apparatus - Google Patents

Description

  The present invention relates to a dropout color identification processing program, method, and apparatus. In particular, the present invention can be used in operations where the dropout color differs for each form in the OCR dropout process.

  A drop that prevents a particular color (dropout color) from appearing as an image when scanning a form data in a ruled line or character color printed or printed in advance on a form. Perform out processing.

  In the conventional dropout processing, a dropout color read-only image reader is used, and one dropout color of the form must be selected in accordance with the light source characteristics of the image reader.

  Therefore, for example, when a plurality of forms having different dropout colors such as a red dropout color receipt slip and a blue dropout color withdrawal slip are to be subjected to a dropout process in advance, red and blue Using an image reader with two light sources, processing was performed by switching the reading light source according to the dropout color of the form to be processed.

  In conventional dropout processing, when the form was read, the dropout color had to be determined in advance, so the slips with different dropout colors for each form or mixed with different dropout colors were processed together. It was impossible to do.

  An object of the present invention is to automatically identify a dropout color of a form to be processed and perform a dropout process based on the identified dropout color.

  It is another object of the present invention to perform dropout processing by automatically identifying each dropout color of a form in which a plurality of dropout colors are used without designating an area.

  According to the present invention, it is possible to perform processing that does not depend on the dropout color, and it is not necessary to identify the dropout color in advance, and processing can be performed on a form in which dropout colors are mixed.

  In order to achieve the above object, a program according to the present invention includes a plane acquisition process for separately generating an RGB plane image from form image data to acquire an R plane, a G plane, and a B plane, and the form. A luminance signal image is generated from the image and further binarized to generate a luminance signal binarized image, and the respective dynamic determination threshold values acquired in the plane acquisition processing are obtained. , Based on the determination threshold, each plane is binarized to generate each plane binarized image, the luminance signal binarized image, and each plane binary For each pixel, a combining process for generating a binary combined image of each plane by combining each pixel, and each plane 2 Calculating a pixel frequency of each composite image, and causing a computer to execute a dropout plane selection process of selecting a dropout plane from the plane that is the basis of each plane binary composite image using the pixel frequency It is.

  The present invention operates a computer as follows. The computer in which the present invention is executed is a plane image obtained by removing each RGB system color component from the form image data, and is obtained by removing the R system color component and the G system color component. A B plane from which the plane and B color components are removed is generated. Then, a density characteristic is obtained for each plane of the R plane, the G plane, and the B plane, and a discrimination threshold value is calculated based on a predetermined formula from the density characteristics of each plane. Based on this threshold value, Each plane of the R plane, the G plane, and the B plane is binarized, and a binarized image of each plane is generated. Further, a luminance signal image is generated from the form image data, the luminance signal image is binarized, and a luminance signal binarized image is generated.

  Further, the corresponding pixels of the luminance signal binarized image and the binarized image of each plane are combined to generate a binary combined image of each R / G / B plane. Then, the pixel frequency (black pixel frequency) is calculated for each binary composite image of each plane. Based on the calculated pixel frequency, one binary composite image showing a tendency of different pixel frequencies is selected from the R / G / B plane binary composite image and corresponds to the selected binary composite image. The color system removed by the plane to be identified is specified as the dropout color.

  As a result, even when the portion to be dropped out of the form has shades of color printing, the dropout processing can be performed accurately without causing a drop in the performance of the dropout processing.

  As an invention related to the present invention, plane acquisition processing for separately generating and generating R, G, and B planes from RGB image plane images from form image data, the R plane, G plane, Threshold value calculation processing for obtaining density characteristics for each plane of the B plane, and automatically calculating a discrimination threshold value for each plane from the density characteristics, and using the determination threshold value, from each plane There is a program that causes a computer to execute dropout plane selection processing for selecting a dropout plane.

  Further, as an invention related to the present invention, plane acquisition processing means for separately generating and generating R, G, and B planes from RGB image plane images from the report image data, and luminance signals from the report image Luminance signal binarization processing means for generating an image and binarizing to generate a luminance signal binarized image; and determining each of the dynamic determination threshold values acquired in the plane acquisition processing; A plane binarization processing means for binarizing each plane based on a threshold value to generate each plane binarized image, the luminance signal binarized image, and each plane binarized image First plane processing unit for generating a first binary composite image by combining each plane binary image with each pixel, A binary composite image and a single binary image that is not processed by the first composite processing means among the plane binary images are combined for each pixel to generate a second binary composite image. And a second synthesizing processing means.

  In the present invention, after the image data of a form to be dropped out is separated into RGB plane images and luminance signal images (hereinafter referred to as Y signal images), the dropout color is identified by one of the following methods. be able to.

  a) The dropout plane is identified from the histogram of each plane, and the color of the dropout plane is set as the dropout color.

  b) The dropout plane is identified by the black pixel frequency calculated from the composite image of the Y signal and each plane, and the color of the dropout plane is set as the dropout color.

  c) An XOR operation is performed on the Y signal and each plane to identify the dropout plane, and the color of the dropout plane is set as the dropout color.

  A highly accurate dropout process can be performed by a process using the dropout color identified for each form, that is, a composition process using a dropout plane.

  In addition, by repeating or combining the processes a) to c), a highly accurate dropout process can be performed on a form image data in which a plurality of dropout colors are mixed without exchanging the light source. Can do.

  Note that a program for realizing each means, function, or element of the present invention by a computer can be stored in an appropriate recording medium such as a portable medium memory, a semiconductor memory, or a hard disk that can be read by the computer. It is provided by being recorded on a medium, or provided by transmission / reception using various communication networks via a communication interface.

  In the present invention, a plane in which each RGB system color component separated and generated from the acquired form image is removed and generated from the tendency of the pixel frequency of the binarized plane binarized by a predetermined threshold value is used. , A pixel having a different pixel frequency tendency is specified as a dropout color, and a dropout process is performed by synthesizing it with a luminance signal binarized image generated from a form image.

  As a result, it is not necessary to provide a dedicated light source for a predetermined dropout color as in the prior art, and the dropout process can be performed accurately without causing a drop in the performance of the dropout process due to the shade of the printout of the dropout color. Can be done.

  Further, according to the present invention, dropout processing for combining the binary image of each RGB system plane generated separately from the image of the form and the binary image of the luminance signal can be performed by combining each system plane. As a result, it is possible to perform dropout processing for a form in which a plurality of dropout colors are used without determining areas for each dropout color.

  In particular, it is possible to perform dropout processing at one time without storing the area position etc. on the different multidropout color forms before processing, so even when various multidropout color forms are mixed, It becomes possible to perform dropout processing accurately and easily.

Hereinafter, embodiments of the present invention will be described together with related inventions.
[First Embodiment]
FIG. 1 shows a configuration example of an apparatus according to the present invention in the first embodiment.

  The dropout processing apparatus 100 includes an image acquisition unit 101, a shadow correction processing unit 102, an RGB plane division unit 103, an image storage unit 104, a histogram calculation unit 105, a determination threshold value calculation unit 106, a dropout A plane selection unit 107, a dropout plane binarization processing unit 108, a luminance signal generation unit 109, a luminance signal binarization processing unit 110, a binarized image composition processing unit 111, and a blur correction processing unit 112; Is provided.

  The image acquisition unit 101 is means for acquiring the color image data (form image) of the form 2. Here, the image acquisition unit 101 is a two-dimensional color CCD. However, the image acquisition unit 101 is not limited to this as long as it can acquire color image data.

  The shadow correction processing unit 102 is a unit that corrects an unnecessary part such as a shadow generated in the form image captured by the image acquisition unit 101 by a known technique such as shading.

  The RGB plane division unit 103 is a means for separating and generating RGB plane images (R plane, G plane, B plane) from the form image. The R plane is image data from which R (red) colors are removed, the G plane is image data from which G (green) colors are removed, and the B plane is B (blue) colors. It is the removed image data. Note that the RGB plane division unit 103 performs correction processing for aligning the resolutions of the R / G / B planes, if necessary.

  The image storage unit 104 is an area for storing each plane separated and generated by the RGB plane division unit 103.

  The histogram calculation unit 105 is a unit that generates a density value histogram for each plane generated by the RGB plane division unit 103 and digitizes the histogram.

  The determination threshold value calculation unit 106 is a unit that calculates a threshold value for each R / G / B plane based on the numerical value of the histogram calculated by the histogram calculation unit 105. The determination threshold value calculation unit 106 uses an automatic determination threshold value method described later. [References: Noriyuki Otsu, Automatic threshold selection method based on discrimination and least-square criterion, IEICE Transactions, J63-D, vol.4, pp.349-356 (1986)]

  The dropout plane selection unit 107 is means for selecting a dropout plane from the three R / G / B planes based on the processing result of the determination threshold value calculation unit 106. For example, in a plane of the same color system as the dropout color, the density value of the portion such as the border of the dropout color is close to the density value of the background color, compared to the other two planes that are not the same color system as the dropout color. Since the threshold values calculated by the judgment threshold value calculation unit 106 are remarkably different values, planes indicating different threshold values are set as dropout planes.

  The dropout plane binarization processing unit 108 is a unit that receives the determination result from the dropout plane selection unit 107 and binarizes the dropout plane. Since the dropout color is removed from the dropout plane, the binarization process makes information other than the dropout color such as black handwritten characters and guide characters.

  The luminance signal generation unit 109 is a unit that generates a Y signal image from each of the R, G, and B planes. The Y signal image may be generated directly from the image of the form 2 acquired by the image acquisition unit 101.

  The luminance signal binarization processing unit 110 is means for binarizing the Y signal image. The luminance signal binarization processing unit 110 performs a local binarization processing method in which a threshold value is set for a local region in order to suppress missing information. The luminance signal binarization processing unit 110 binarizes everything except the background portion as information.

  The binarized image composition processing unit 111 outputs the dropout plane binarized image output from the dropout plane binarization processing unit 108 and the Y signal binarized image output from the luminance signal binarization processing unit 110. , And a means for combining by AND operation. By combining processing, it is possible to output an image from which the dropout color portion is removed.

  The blur correction processing unit 112 is a unit that performs blur correction of an information part such as a handwritten character or blank complement for the binary composite image output from the binarized image synthesis processing unit 111.

  Hereinafter, the processing of the invention in this embodiment will be described in detail.

  FIG. 2 shows a part of the form 2 image captured by the image acquisition unit 101. The form 2 read by the dropout processing apparatus 100 is one in which handwritten characters are written by printing a border line with a red dropout color on a white background.

  The image acquisition unit 101 (color CCD) captures the form 2 and passes the acquired image of the form 2 to the shadow correction processing unit 102. The shadow correction processing unit 102 performs shadow correction on the image of the form 2 using a known shading technique or the like. Then, the RGB plane division unit 103 separates and generates the R plane, the G plane, and the B plane from the corrected form 2 image, and stores these in the image storage unit 104.

  The histogram calculation unit 105 generates a histogram for each of the R plane, the G plane, and the B plane, calculates the generated histogram, and stores numerical values in a table (not shown).

  FIG. 3 shows an example of the histogram of each plane. 3A shows an example of a histogram for the R plane, FIG. 3B shows an example of a histogram for the G plane, and FIG. 3C shows an example of a histogram for the B plane. In this embodiment, since the form 2 is a red-type dropout color form, a drop occurs in the R plane. Therefore, the dropout color portion is the same as the background portion in the histogram of the R plane shown in FIG. Concentration value. On the other hand, as shown in FIGS. 3B and 3C, in the histogram of the B plane or the G plane, a red-type dropout color appears, so that a valley is formed in the waveform of the background portion.

  When the dropout color is a single color (only red) as in the form 2 of this embodiment, the waveform for the white region of the background is different between the dropout color plane and the non-dropout color plane. .

  Therefore, the determination threshold value calculation unit 106 determines a dropout plane by using an automatic determination threshold value method for identifying and determining a portion where the waveform of the histogram is different for each plane.

  The automatic discrimination threshold method is a method of selecting an appropriate threshold for binarization that separates a grayscale image into a target region and a background. In the automatic discrimination threshold method, the threshold value is obtained as follows.

First, a normalized histogram is obtained by setting the total number of pixels for each plane as N = N0 + N1 + N2 + N3... N255. That is, for each density level ni, pi = ni / N (ni: distribution of each density, N: total number of pixels)
A normalized histogram pi is obtained. Then, the average luminance level μt is obtained from the normalized histogram pi by the following formula.

μt = Σi pi (i: 0 to 255)
(I: concentration distribution, pi: normalized histogram)
Next, from pi, μt, a cumulative histogram k is obtained from the following equation.

ω (k) = Σpi (i: from 0 to k)
m (k) = Σi pi (i: from 0 to k)
Next, the interclass variance is obtained from the cumulative histograms ω (k) and m (k) of the normalized histogram by the following formula.

σb = ω (I) (1.0−ω (I))
σ (I) = ((μtω (I) ² −m (I))) / σb (I)
Then, the largest value in the concentration distribution of 0 to 255 in σ (I) is taken out as a discrimination threshold value.

  When calculated by the automatic discrimination threshold method, the R plane threshold Rs = 187, the G plane threshold Gs = 152, and the B plane threshold Bs = 152. The threshold value Rs is the highest and is larger than the threshold values Bs and Gs for the B plane and the G plane.

  In this way, if a valley is formed at a certain high gradation level, it can be estimated as a dropout color. Therefore, there is no valley in the histogram of one plane, and there is a valley in the histogram of another plane. In some cases, a high threshold plane without a valley can be identified as a dropout plane.

  Therefore, the dropout plane selection unit 107 selects an R plane having a higher threshold than the other two planes as a dropout plane, and notifies the dropout plane binarization processing unit 108 of the determination result.

  The dropout plane binarization processing unit 108 binarizes the dropout plane (R plane) on the basis of the threshold value calculated by the determination threshold value calculation unit 106 to provide a binary value for the dropout plane (R plane). Generate an image. FIG. 4 shows an example of the binary image of the dropout plane.

  The luminance signal generation unit 109 generates a Y signal image from each R / G / B plane. The luminance signal binarization processing unit 110 binarizes the Y signal image by the local binarization method.

  Niblack method binarization processing is performed as a local binarization processing method. The Niblack method is a binarization method in which peripheral information (3 × 3, 7 × 7, etc.) is set for a pixel to be obtained and a threshold value is obtained for each pixel.

First, as background separation processing, a rectangular filter (filter size is arbitrary) is set, and the background portion or the information portion is discriminated from the density dispersion value in the filter. Then, a background separation threshold (σmin) is set, and if the calculation result in the filter is greater than or equal to σmin, it is determined that there is information. Assuming that the background separation threshold σmin = 100,
Formula σmin <g2 = E2− (E × E)
(E: Average density in filter, E2: Average density in filter)
If so, it is determined that it is not background. Then, as a threshold value calculation process, if the variance value is equal to or greater than the variance threshold value (σmin), a threshold value of density for the pixel of interest k is obtained to determine whether it is a white pixel or a black pixel. [Reference: W. Niblack, "An Introduction to Digital Image Processing", pp. 115-116, Englewood Cliffs, NJ: Prentice Hall (1986)]
In this embodiment, for example, a 7 × 7 filter is used as the luminance signal binarization processing unit 110 in order to generate an image such as a handwritten character recognition image for OCR. This is because the stroke width of handwritten characters is about 3 dots, so a 7 × 7 filter is desirable.

  FIG. 5 shows an example of local binarization processing using a 7 × 7 filter. First, a rectangular region is set for each pixel of interest k for which a threshold value is to be obtained, three pixels in the vertical and horizontal directions. In the 7 × 7 filter, if the pixel of interest k is the background (in the filter and only the background), it is dropped with the following background separation threshold σmin.

Background separation threshold σmin = 100 (example of this embodiment)
σmin <g2 = E2- (E × E)
E: Density within filter, E2: Squared density within filter Also, since the absolute value of density is not observed, it is also considered as a background when the information part (character stroke, etc.) is completely covered in the 7x7 filter. However, it is possible to perform black interpolation in the character stroke portion by determining the density value.

  When the information part (character stroke) and the background are included in the 7 × 7 filter, it becomes larger than the σmin threshold value. Then, the obtained √g2 is multiplied by a coefficient (−0.1). The effect of suppressing noise around the character can be obtained by setting the under threshold to the obtained threshold. Then, the following threshold value S of the target pixel k is obtained.

Threshold value S of the target pixel k = E + (− 0.1) × √g2
If the pixel of interest k has a density equal to or lower than the obtained threshold value, it becomes a black pixel, and if it is higher, it becomes a white pixel.

  The luminance signal binarization processing unit 110 is an automatic discrimination threshold method used in the dropout plane binarization processing unit 108 in order to estimate an intermediate density level of the luminance signal that is a range in which local binarization processing is performed. Apply. As shown in FIG. 6A, a dynamic two-tone separation threshold value is obtained for the entire luminance signal image by the automatic discrimination threshold method, and the direction of low density (density for black) is centered on the threshold value. For the higher density (white density) direction, the ratio is set by the following formula to calculate the intermediate density level partial estimation area.

Intermediate density level range coefficient a = 0.10 to 0.90,
Intermediate density level range coefficient b = 0.10 to 0.90,
Black area density average b_ave
= Density average of density distribution 0 to density distribution ot Density average of white area w_ave
= Concentration average of density distribution ot + 1 to density distribution 255 Lower limit bt of intermediate density level
= ((1.0-a) * ot) + (a * b_ave)
Upper limit of intermediate concentration level wt
= ((1.0-b) * ot) + (b * w_ave)
Only the estimated area (bt to wt) is binarized by the local binarization method, and the other b_ave areas are black and the w_ave areas are white. As a result, significant high-speed processing is possible. The ratio of the coefficients a and b has a lot of local binarization. However, if the ratio is increased too much, it becomes useless processing when binarizing a handwritten character.

  In this embodiment, the local binarization is performed in the range of 100 by setting the coefficient a = 0.35 and the coefficient b = 0.62, that is, by performing the local binarization processing on the entire range of 5 to 8%. We were able to obtain the same level of results as we did. Since the optimum ratio varies depending on the number of characters entered in the form, the coefficient is set as appropriate.

  FIG. 6B shows an example of a Y signal binary image by the local binarization processing method. In the Y signal binary image, everything except the background is used as an information portion (black pixel).

  After that, the binarized image composition processing unit 111 performs composition processing on the dropout plane (R plane) binary image and the Y signal binary image by AND operation to generate a new YR binary image.

  FIG. 7 shows an example of a YR binary image. The red dropout color information of the Y signal binary image is binarized and becomes information, but in the dropout plane (R plane) binary image, the red dropout color is removed and is not information. Therefore, the dropout color having red characteristics is also removed from the image synthesized by the AND operation.

  Thereafter, the blur correction processing unit 112 may cause a blur in an information unit such as a handwritten character or a guide during the processing of the binary image synthesis processing unit 111. Perform correction processing.

  In the AND operation in the binarized image composition processing unit 111, if neither the dropout plane binary image nor the Y signal binary image is output as binarized information, the information is displayed on the composite image. Is not output as. For this reason, a portion that is not output as an information portion on one side causes blurring. As a cause of blurring, it can be considered that the gradation level used as information is high (close to the gradation of the background portion).

  The blur correction processing unit 112 performs one of the following correction processes.

(A) First blur correction process of Y signal image When the gradation of the blur part is close to the gradation of the background level, it is processed as the background in the local binarization method of the luminance signal binarization processing unit 110. It may end up. In the local binarization method, background separation processing is performed as preprocessing, and a binarization threshold value is obtained only at locations where information exists. In this background separation processing, the luminance signal binarization processing unit 110 locally filters the image of the form 2 to obtain a background level gradation variance, and if there is no threshold value (density fluctuation), the background It will be judged as a part (white).

  Therefore, the blur correction processing unit 112 improves the blur by performing correction that lowers the background separation threshold of the luminance signal binarization processing unit 110.

  FIG. 8 shows a result example based on the presence or absence of correction processing by the blur correction processing unit 112. FIG. 8A shows an example of a binarized image portion based on the background separation threshold calculated by the luminance signal binarization processing unit 110. FIG. 8B shows an example of a binarized image portion binarized by the luminance signal binarization processing unit 110 using the background separation threshold set lower by the blur correction processing unit 112. In the image of FIG. 8B, it can be seen that the blur is improved compared to FIG. 8A.

(B) Second blur correction process of Y signal image In the luminance signal binarization processing unit 110, the local binarization process is limited to a certain range. As shown in FIG. 9, a predetermined gradation range (bt to wt) is obtained from the histogram of the Y signal image, and processing is performed only within this range to improve the processing efficiency. Therefore, even when an information part such as a handwritten character is out of this gradation range, blurring still occurs. In other words, if there is an information part for a gradation above the range bt in FIG. 9, the information is unconditionally determined to be white in terms of processing logic.

  Therefore, the blur correction processing unit 112 corrects the blur by correcting the values of the coefficients a and b so that the luminance signal binarization processing unit 110 expands the gradation range (bt to wt) that is the target of local binarization. To improve.

  FIG. 10 shows a result example based on the presence or absence of correction processing by the blur correction processing unit 112. FIG. 10A shows an example of a binarized image portion processed in the gradation range calculated by the luminance signal binarization processing unit 110. FIG. 10B is an example of a binarized image portion processed in the gradation range expanded by the blur correction processing unit 112. In the image of FIG. 10B, it can be seen whether the blur is improved as compared with FIG.

(C) Blur correction processing of dropout plane image Even if an image without blur is generated from the Y signal binary image by the above blur correction processing, the other image to be synthesized, information by the dropout plane binary image If it does not appear, a blur will occur. Therefore, the blur correction must be performed so that information is also output for the binary image of the dropout plane.

  Since the dropout color binary image only needs to output information other than the dropout color, the dropout plane binarization processing unit 108 only needs to be able to binarize to the extent that there is no background noise. In addition, the determination threshold value calculation unit 106 tends to obtain a threshold value that becomes slightly thinner because the automatic determination threshold value method is used.

  Therefore, in order to obtain a certain degree of darkness in the dropout plane, that is, in order for the binarized image composition processing unit 111 to binarize the dropout plane more widely than the Y signal binary image, the blur correction processing unit 112 The threshold value is corrected by the offset with respect to the threshold value actually calculated by the determination threshold value calculation unit 106.

  The blur correction processing unit 112 increases the ratio by an arbitrary value (preferably 30 to 40%) with respect to the threshold value calculated by the determination threshold value calculation unit 106. In this embodiment, it is obtained from the evaluation result that the optimum value is obtained by multiplying the threshold calculated by the determination threshold calculation unit 106 by 1.35. As shown in FIG. 11, the blur correction processing unit 112 sets the threshold value 171 by 1.35 times the threshold value 127 calculated by the determination threshold value calculation unit 106. FIG. 12 shows an example of an output image depending on whether or not the blur correction process is performed. FIG. 12A is an example of an image when the determination threshold is not corrected, and FIG. 12B is an example of an image that is corrected to increase the determination threshold by 1.35 times. is there.

  By performing the blur correction processes A) to C) alone or in combination, blurring that occurs when the binarized image synthesis processing unit 111 performs the synthesis process can be avoided. Then, the blur correction processing unit 112 outputs an image 3 as shown in FIG.

  FIG. 13 shows a process flow of the dropout plane selection process in the first embodiment. It is assumed that R / G / B planes are separately generated by the RGB plane dividing unit 103 before the selection process.

  First, the histogram calculation unit 105 acquires and digitizes a histogram for each of R, G, and B planes, and stores the numerical value in a table (step S1). Then, the determination threshold value calculation unit 106 calculates the automatic threshold values Rs, Gs, and Bs of each plane using the digitized histogram as an argument (step S2). The dropout plane selection unit 107 selects a dropout plane by the following process. That is, if the threshold value Rs is larger than Gs (step S3) and Rs is larger than Bs (step S4), the R plane is selected as a dropout plane (step S5).

  If the threshold value Rs is not larger than Gs (step S3) and Gs is larger than Bs (step S6), the G plane is selected as a dropout plane (step S7).

  Further, if the threshold value Rs is not larger than Gs (step S3) and further Rs is not larger than Bs (step S4), or if Gs is not larger than Bs (step S6), the B plane is set as a dropout plane. Select (step S8).

[Second Embodiment]
In the second embodiment, the dropout processing device 200 identifies the dropout plane by calculating the black pixel frequency from the composite image of the Y signal image and each plane image. Then, an image image 3 such as a handwritten character is output using the identified dropout plane.

  FIG. 14 shows a configuration example of an apparatus according to the present invention in the second embodiment.

  The dropout processing device 200 includes an image acquisition unit 101, a shadow correction processing unit 102, an RGB plane division unit 103, an image storage unit 104, a luminance signal generation unit 109, a luminance signal binarization processing unit 110, Plane threshold value calculation units 201, 202, 203 for each of the R, G, B planes, binarization processing units 204, 205, 206 for each of the R, G, B planes, and three AND operation units 207, 208, 209, a black pixel frequency calculation unit 210, a binarized image composition processing unit 211, and a blur correction processing unit 212.

  Among the processing units of the dropout processing device 200, the image acquisition unit 101, the shadow correction processing unit 102, the RGB plane division unit 103, the image storage unit 104, the luminance signal generation unit 109, and the luminance signal binarization. The processing unit 110 performs the same processing as each processing unit to which the same number constituting the dropout processing apparatus 100 of FIG. 1 is assigned.

  The R plane threshold value calculation unit 201 is means for calculating the threshold value of the R plane image using the automatic discrimination threshold method. Similar to the R-plane threshold value calculation unit 201, the G-plane threshold value calculation unit 202 and the B-plane threshold value calculation unit 203 are means for calculating dynamic threshold values for the G plane and the B plane, respectively.

  The R plane binarization processing unit 204 is a unit that binarizes the R plane based on the threshold value calculated by the R plane threshold value calculation unit 201 and outputs an R plane binary image. Similarly to the R plane binarization processing unit 204, the G plane binarization processing unit 205 and the B plane binarization processing unit 206 are calculated by the plane threshold calculation units 202 and 203 for the G plane and the B plane. This is a means for binarizing the G plane and B plane based on the threshold value.

  The AND operation unit 207 is a means for combining the R plane binary image and the Y signal binary image by AND operation. Similarly, the AND operation unit 208 is a means for combining the G plane binary image and the Y signal binary image by AND operation, and the AND operation unit 209 performs the B plane binary image and the Y signal binary image. It is a means for performing composition processing by AND operation.

  The black pixel frequency calculation unit 210 is a unit that calculates the black pixel frequency of the combined binary image synthesized by the AND operation units 207, 208, and 209, and selects a dropout plane based on the black pixel frequency.

  The binarized image synthesis processing unit 211 is a means for synthesizing the selected dropout plane and the Y signal binary image by an AND operation.

  The blur correction processing unit 212 is a unit that performs blur correction on the binarized composite image synthesized by the binarized image synthesis processing unit 211.

  Processing in this embodiment will be described. The processing from the image acquisition unit 101 to the RGB plane division unit 103 and the processing of the luminance signal generation unit 109 and the luminance signal binarization processing unit 110 are the same as those in the first embodiment.

  The plane threshold value calculation units 201, 202, and 203 for each R / G / B plane each set the dynamic threshold value of each plane in substantially the same manner as the determination threshold value calculation unit 106 in the first embodiment. Calculate using the automatic discrimination threshold method.

  Thereafter, the plane binarization processing units 204, 205, and 206 for each R / G / B plane are calculated for each plane in substantially the same manner as the dropout binarization processing unit 108 in the first embodiment. Each plane is binarized based on the threshold value.

  Next, the AND operation unit 207 combines the R plane binary image and the Y signal binary image by an AND operation to generate a new YR binary image. Similarly, the AND operation unit 208 and the AND operation unit 209 combine the plain binary image and the Y signal binary image by an AND operation to generate a YG binary image and a YB binary image, respectively.

  FIG. 15 shows an example of an image generated by combining processing by the AND operation units 207, 208, and 209. It can be seen that by combining processing, only information (black) in both the Y signal binary image and the binary image of each plane is information of a new binary image synthesized.

  FIG. 15A shows an example of a YR binary image obtained by combining a Y signal binary image and an R plane binary image, and FIG. 15B shows a combination of a G plane binary image and a Y signal binary image. FIG. 15C shows an example of a YG binary image in which a B plane binary image and a Y signal binary image are combined.

  Next, the black pixel frequency calculation unit 210 checks the black pixel frequency which is the information amount of each of the synthesized YR binary image, YG binary image, and YB binary image.

  The black image frequency should be different between the composite image based on the dropped-out plane and the composite image that is not, and the composite image based on the dropped-out plane is compared with other composite images. The black pixel frequency should be low.

  In the YR binary image shown in FIG. 15A, all red dropout colors disappear, so only information (black pixels) such as characters and guides remains. The black pixel frequency YRP of the YR binary image calculated by the black pixel frequency calculation unit 210 is 11.823%. Further, in the YG plane binary image of FIG. 15B, the dropout color portion remains and the character frame of the dropout color is output as information. Therefore, the black pixel frequency YGP is increased to 13.653%. Further, in the YB binary image of FIG. 15C, the black pixel frequency YBP is increased to 13.787% as in the YG binary image.

  Further, assuming that the total number of pixels of the form 2 is 3,115,580 pixels, in the case of each binary image in FIG. 5, the difference in black pixel frequency between the YR binary image and the YG binary image is 1.830% (about 57 , 015 pixels), and the difference in black pixel frequency between the YR binary image and the YB binary image is 1.964% (about 61,189 pixels). On the other hand, the difference in the black pixel frequency between the YG binary image and the YB binary image that are not dropped out is 0.134% (approximately 4,174 pixels), and only 0.1% (approximately 3000 pixels) is recognized. I can't.

  As described above, the black pixel frequency in the synthesized binary image differs between the plane that is dropped out and the plane that is not dropped out. Further, the black pixel frequencies of the binary images synthesized on the plane that does not drop out are substantially the same.

  Accordingly, the black pixel frequency calculation unit 210 can determine that the plane that is the source of the binary image that differs in value from the black pixel frequency of the binary image synthesized from the other two planes is a dropped-out plane. The R plane that is the basis of the binary image with the lowest black pixel frequency is determined as the dropout plane.

  The binarized image composition processing unit 211 receives the determination result from the black pixel frequency calculation unit 210, and, similarly to the binarized image composition processing unit 111 of the first embodiment, the R-plane binary image and the Y image The signal binary image is combined with an AND operation.

  In addition, the blur correction processing unit 212 performs correction processing combining any one or several of the correction processes, similar to the blur correction processing unit 112 of the first embodiment.

  FIG. 16 shows a processing flow of dropout plane selection processing in the second embodiment. Similarly to the processing flow shown in FIG. 13, it is assumed that R / G / B planes are generated separately before the selection processing.

  First, the luminance signal image binarization processing unit 110 locally binarizes the Y signal image generated by the luminance signal generation unit 109 (step S11). Then, the processing from step S12 to step S15 is performed for each plane of R / G / B.

  First, the plane threshold value calculation unit 201, 202, 203 for each plane acquires and digitizes a histogram for each plane (step S12), and uses the digitized histogram as an argument as an automatic threshold value (Rs, Gs, Bs). ) Is calculated (step S13), and the plane binarization processing units 204, 205, and 206 of each plane binarize each plane based on the automatic threshold value (step S14). Then, the AND operation units 207, 208, and 209 perform an AND operation on the Y signal binary image and each plane binary image to generate a composite binary image (step S15).

  Then, the black pixel frequency calculation unit 210 calculates the black pixel frequencies YRP, YGP, and YBP of each synthesized binary image (step S16).

YR binary plane black pixel frequency YRP
= (Number of black pixels in YR binary value ÷ total number of pixels) × 100
YG binary plane black pixel frequency YGP
= (Number of black pixels of YG binary value ÷ total number of pixels) × 100
Black pixel frequency YBP of YB binary plane
= (Number of black pixels in YB binary value ÷ total number of pixels) × 100
In addition, the dropout plane is selected by the following processing. That is, when YRP is smaller than YGP (step S17) and YRP is smaller than YGP (step S18), the R plane is selected as the dropout plane (step S19).

  If YRP is not smaller than YGP (step S17) and YGP is smaller than YBP (step S20), the G plane is selected as the dropout plane (step S21).

  If YRP is smaller than YGP (step S17) and YRP is not smaller than YGP (step S18), or if YGP is not smaller than YBP (step S20), the B plane is selected as the dropout plane (step S20). S22).

[Third Embodiment]
In the third embodiment, the dropout processing device 300 identifies a dropout plane by performing an XOR operation on the Y signal image and each RGB plane image and extracting a remaining portion of the dropout color. Then, an image image 3 such as a handwritten character is output using the identified dropout plane.

  FIG. 17 shows a configuration example of an apparatus according to the present invention in the third embodiment.

  The dropout processing device 300 includes an image acquisition unit 101, a shadow correction processing unit 102, an RGB plane division unit 103, an image storage unit 104, a luminance signal generation unit 109, a luminance signal binarization processing unit 110, Plane threshold value calculation units 201, 202, and 203 for each of the R, G, and B planes, binarization processing units 204, 205, and 206 for each of the R, G, and B planes, and three XOR operation units 301, 302, 303, a black pixel frequency calculation unit 304, a binarized image composition processing unit 305, and a blur correction processing unit 306.

Among the processes of the dropout processing device 3 00, an image acquisition unit 101, a shadow correction processing unit 102, an RGB plane dividing unit 103, an image storage unit 104, a luminance signal generation unit 109, the luminance signal binary The processing unit 110 performs the same processing as each processing unit assigned the same number that constitutes the dropout processing device 100 of FIG. and 202, 203, R, G, and binarization processing unit 204, 205, and 206 for each B plane, the same as the processing units of the same number is applied which constitutes a drop-out processing unit 200 of FIG. 1 4 Process.

  The XOR operation unit 301 is a unit that synthesizes the R-plane binary image and the Y signal binary image generated by the luminance signal binarization processing unit 110 by an XOR operation. Similarly, the XOR operation unit 302 is a means for combining the G plane binary image and the Y signal binary image by XOR operation, and the XOR operation unit 303 combines the B plane binary image and the Y signal binary image. This is a means for synthesizing by XOR operation.

  The black pixel frequency calculation unit 304 calculates the dropout color remaining frequency of the composite binary image synthesized by the XOR operation units 301, 302, and 303, and selects a dropout plane based on the remaining frequency. It is means to do.

  The binarized image composition processing unit 305 is means for performing composition processing on the selected dropout plane and the Y signal binary image by AND operation.

  The blur correction processing unit 306 is means for performing blur correction on the binarized composite image synthesized by the binarized image synthesis processing unit 305.

  Processing in this embodiment will be described.

  The processing from the image acquisition unit 101 to the RGB plane division unit 103 and the processing of the luminance signal generation unit 109 and the luminance signal binarization processing unit 110 are the same as those in the first embodiment.

  Also, the plane threshold value calculation units 201, 202, and 203 for each R / G / B plane obtain a dynamic threshold value for each plane using a local binarization method. Then, the plane binarization processing units 204, 205, and 206 perform binarization processing using threshold values obtained for the respective R / G / B planes.

  Next, the XOR operation unit 301 performs composition processing by XOR operation on the R plane binary image and the Y signal binary image to generate a new YRX binary image. Similarly, the XOR operation unit 302 combines the G plane binary image and the Y signal binary image to generate a YGX binary image, and the XOR operation unit 303 generates the B plane binary image and the Y signal binary image. A YBX binary image is generated by combining the images. FIG. 18 shows an example of a binary image synthesized by the XOR operation units 301, 302, and 303. 18A shows a YRX binary image example, FIG. 18B shows a YGX binary image example, and FIG. 18C shows a YBX binary image example.

  Thereafter, the black pixel frequency calculation unit 304 checks the remaining frequency of the dropout color (black pixel) of each binary image processed by the XOR operation units 301, 302, and 303.

  Usually, since all information is binarized in the Y signal binary image, XOR operation is performed with the binary image of the plane that is being dropped out, so that the dropout color portion is information (black pixels). As a result, an image in which the dropout color part is raised is generated.

  Therefore, the black pixel frequency calculation unit 304 has the highest number of information among the three images of the YRX binary image, the YGX binary image, and the YBX binary image, that is, the binary image having the highest remaining frequency of the dropout color portion. Select the base image as the dropout plane.

  The remaining frequency YRXP of the YRX binary image in FIG. 18A is 5.943% (6.014% after the blur correction process), and the remaining frequency of the YGX binary image shown in FIG. 18B is 4.424. % (3.577% after blur correction processing), and the remaining number frequency YBXP of the YBX binary image in FIG. 18C was 4.404% (4.219% after blur correction processing).

  Further, the difference rate of the remaining number frequency was 1.519% for the YRX binary image and the YGX binary image, and 1.539% for the YRX binary image and the YBX binary image. On the other hand, the difference rate of the remaining frequency between the YGX binary image and the YBX binary image was 0.020%.

  In the binary plane synthesized based on the plane that is dropped out, the amount of information (number of black pixels) corresponding to the dropout color increases as a result of the XOR bit operation. Therefore, the black pixel frequency calculation unit 304 can identify the dropout plane by setting the difference rate threshold value and the spread of the difference rate between other planes. In this embodiment, the black pixel frequency calculation unit 304 identifies the R plane as a dropout plane.

  FIG. 19 shows a processing flow of dropout plane selection processing in the third embodiment. Similarly to the processing flow shown in FIG. 13, it is assumed that R / G / B planes are generated separately before the selection processing.

  The processing from step S31 to step S34 is the same as the processing from step S11 to step S14 in the processing flow shown in FIG.

  The XOR operation units 301, 302, and 303 generate a composite binary image by performing an XOR operation on the Y signal binary image and each plane binary image (step S35).

  Then, the black pixel frequency calculation unit 304 calculates the remaining frequency YRXP, YGXP, YBXP of the dropout color portion of each synthesized binary image (step S36).

YR X Binary Plane Remaining Frequency YRXP
= (YR X binary black pixel number / total number of pixels) × 100
YG X Binary Plane Remaining Frequency YGXP
= (YG X binary black pixel number / total number of pixels) × 100
Remaining number frequency YBXP of YB X 2 value plane
= (YB X binary black pixel number / total number of pixels) × 100
In addition, the dropout plane is selected by the following processing. That is, when YRXP is larger than YGXP (step S37) and YRXP is larger than YGXP (step S38), the R plane is selected as the dropout plane (step S39).

  If YRXP is not larger than YGXP (step S37) and YGXP is larger than YBXP (step S40), the G plane is selected as the dropout plane (step S41).

  If YRXP is larger than YGXP (step S37) and YRXP is not larger than YGXP (step S38), or if YGXP is not larger than YBXP (step S40), the B plane is selected as the dropout plane (step S40). S42).

[Fourth Embodiment]
In the dropout processing apparatuses of the first to third embodiments, the dropout color of the read form may not be identified. For example, the dropout color may not be determined when there is no valley between the background portion of the histogram of each RGB plane and the dropout color or when the printout frequency of the dropout color is low. Also, the dropout color cannot be determined for a form having a plurality of dropout colors.

  In the fourth embodiment, the dropout processing device 400 combines a variety of methods for identifying the dropout color described in the above embodiments to combine a form or a plurality of drops that cannot identify the color system of the dropout color. In the case of a form in which outcolors are mixed, a desired image image 3 or 4 is output by performing dropout processing with a plurality of dropout colors on the image of one form 2.

  FIG. 20 shows a part of the image of the form 2 to be processed in this embodiment. Form 2 processed in this embodiment is a form in which a seal frame is printed with a blue dropout color, an entry character frame is printed with a red dropout color, and black handwritten characters are written. Is a mixed form.

  FIG. 21 shows a configuration example of an apparatus according to the present invention in the fourth embodiment.

  The dropout processing device 400 includes an image acquisition unit 101, a shadow correction processing unit 102, an RGB plane division unit 103, an image storage unit 104, a histogram calculation unit 105, a determination threshold value calculation unit 106, a dropout A plane selection unit 107, a dropout plane binarization unit 108, a luminance signal generation unit 109, a luminance signal binarization processing unit 110, a binary image composition processing unit 111, a blur correction processing unit 112, Plane threshold value calculation units 201, 202, and 203 for each of the R, G, and B planes, binarization processing units 204, 205, and 206 for each of the R, G, and B planes, a YR composition calculation unit 401, and a YG composition A calculation unit 402, a YB composition calculation unit 403, and a blur correction processing unit 404 are provided.

  Among the processing units of the dropout processing device 400, the image acquisition unit 101, the shadow correction processing unit 102, the RGB plane division unit 103, the image storage unit 104, the histogram calculation unit 105, and the determination threshold value calculation unit. 106, a dropout plane selection unit 107, a dropout plane binarization unit 108, a luminance signal generation unit 109, a luminance signal binarization processing unit 110, a binary image composition processing unit 111, and a blur correction. The processing unit 112 performs substantially the same processing as each processing unit to which the same number that constitutes the dropout processing apparatus 100 of FIG. 1 is assigned, and a plane threshold value calculation unit 201 for each of the R, G, and B planes, 202 and 203 and the binarization processing units 204, 205, and 206 for each of the R, G, and B planes constitute the dropout processing device 200 of FIG. Perform substantially the same processing as the processing unit to which an ID is assigned.

  The YR synthesis operation unit 401 performs synthesis processing on the Y image binary image or the binary image synthesized by the other synthesis operation units 402 and 403 and the R plane binary image by AND operation, and creates a new binary synthesis image. Is a means for generating

  The YG synthesis calculation unit 402 performs a synthesis process on the Y image binary image or the binary image synthesized by the other synthesis calculation units 401 and 403 and the G plane binary image by an AND operation to obtain a new binary synthesis image. Is a means for generating

  The YB synthesis operation unit 403 performs an AND operation on the Y signal binary image or the binary image synthesized by the other synthesis operation units 401 and 402 and the B plane binary image, and performs a new binary synthesis image. Is a means for generating

  Of these combination calculation units 401, 402, and 403, two combination calculation units may be combined for processing, or all three combination calculation units may be processed. As for the processing order, processing may be performed from any synthesis operation unit, and the same result is obtained.

  When the dropout plane can be determined based on the threshold calculated by the determination threshold calculation unit 106, the dropout plane selection unit 107 notifies the determination result to the dropout plane binarization processing unit 108. To do. On the other hand, when the dropout plane cannot be determined, the process is passed to the plane threshold value calculation units 201, 202, and 202 of each plane of R / G / B. In this case, all the plane threshold value calculation units may be activated, or only a predetermined plane threshold value calculation unit may be activated, or in a predetermined order. You may make it start a plane threshold value calculation part one by one.

  As described above, the luminance signal binarization processing unit 110 uses the local binarization method for the Y signal image generated by the luminance signal generation unit 109 so that all Y signal images including the dropout color are included. The optimal binarization is performed so as to output the information. FIG. 22A shows an example of a Y signal binary image.

  In this embodiment, an example of processing the red and blue dropout colors of the image of the form 2 shown in FIG. 21 will be described.

  The R plane threshold value calculation unit 201 receives the notification from the dropout plane selection unit 107 and calculates the threshold value of the R plane using the automatic discrimination threshold method. Then, the R-plane binarization processing unit 204 performs either a simple binarization process or an optimal binarization process by a local binarization method using the calculated threshold value. An R-plane binary image shown in (B) is generated.

  Then, the YR synthesis calculation unit 401 combines the Y signal binary image of FIG. 22A and the R plane binary image of FIG. 22B by AND operation to newly generate a YR binary image.

  FIG. 22C shows an example of a YR binary image. The red system dropout color part of the Y signal binary image is binarized and output as information, but the red system dropout color part is not output as information in the R plane binary image, so AND operation processing As a result, the red color dropout color portion is erased without being output as information.

  The B plane threshold value calculation unit 203 receives the notification from the dropout plane selection unit 107 and calculates a threshold value for the B plane using the automatic discrimination threshold method. Then, the B plane binarization processing unit 206 performs either a simple binarization process or an optimal binarization process by a local binarization method using the calculated threshold value, and FIG. A B plane binary image shown in (D) is generated.

  After that, the YB composition calculation unit 403 combines the YR binary image generated by the YR composition calculation unit 401 and the B plane binary image by AND operation to newly generate a YRB binary image.

  FIG. 22E shows an example of a YRB binary image. In the YRB binary image, the red dropout color portion is already deleted, but the blue dropout color portion is binarized and output as information. On the other hand, in the B plane binary image, the blue dropout color portion is not output as information, and therefore, the blue characteristic dropout color is deleted without being output as information by the AND operation processing.

  Therefore, at this point, the two dropout color portions having different red characteristics and blue characteristics are erased.

  If the type of the dropout color is not known, the YG composition calculation unit 402 further masks (combines) the G-plane binary image with the YRB binary image, thereby obtaining a green characteristic (G system). The dropout color part of can also be erased.

  As mentioned above, although this invention was demonstrated by the embodiment, this invention can be variously deformed in the range of the main point.

  For example, in the embodiment of the present invention, the configuration in which the dropout processing devices 100, 200, 300, and 400 according to the present invention include the image acquisition unit 101 that reads a color image from the form 2 and the shadow correction processing unit 102 has been described. The dropout processing devices 100, 200, 300, and 400 may input the color image data of the form instead of the form 2 via various media or transmission paths. In this case, the image acquisition unit 101 and the shadow correction processing unit 102 are unnecessary.

  In addition, each processing unit that binarizes various images constituting the apparatus according to the present invention performs binarization processing by combining a global binarization method and a local binarization method as a binarization method. It may be.

  Also, the blur correction processing units 112, 212, 306, and 404 guarantee the quality of the output image and may be unnecessary.

  Further, the dropout processing apparatuses 100, 200, 300, and 400 according to the present invention can be implemented by being incorporated in a form 2 reading apparatus.

  The features of the embodiments and examples of the present invention are listed as follows.

(Supplementary note 1) A program for causing a computer to execute a process of identifying a dropout color used in a form dropout process,
A plane acquisition process for separately generating and generating R, G, and B planes for each RGB system from the form image data;
A threshold value calculation process for obtaining density characteristics for each of the R plane, G plane, and B plane, and automatically calculating a discrimination threshold value for each plane from the density characteristics;
Dropout plane selection processing for selecting a dropout plane from each of the planes using the determination threshold,
A dropout color identification processing program which is executed by a computer.

(Supplementary note 2) A program for causing a computer to execute a process for identifying a dropout color used in a form dropout process,
A plane acquisition process for separately generating and generating R, G, and B planes for each RGB system from the form image data;
A luminance signal binarization process for generating a luminance signal image from the form image and further binarizing to generate a luminance signal binary image;
Plane 2 for obtaining each of the dynamic determination threshold values acquired in the plane acquisition process, and binarizing each plane based on the determination threshold value to generate each plane binarized image Valuation processing,
A combining process for generating a binary combined image of each plane by combining the luminance signal binarized image and each of the plane binarized images for each pixel;
A pixel frequency of each of the plane binary composite images, and a dropout plane selection process for selecting a dropout plane from the plane that is the basis of each plane binary composite image using the pixel frequencies.
A dropout color identification processing program which is executed by a computer.

(Appendix 3) In the dropout color identification processing program described in Appendix 2,
In the synthesis process, an AND operation is performed for each pixel to generate each plane binary synthesized image;
In the dropout plane identification process, the black pixel frequency of each plane binary composite image is calculated, and the plane that is the basis of the plane composite binary image having the minimum black pixel frequency is selected as the dropout plane. And
A dropout color identification processing program which is executed by a computer.

(Appendix 4) In the dropout color identification processing program described in Appendix 2,
In the synthesis process, an XOR operation is performed for each pixel to generate each plane binary synthesized image;
In the dropout plane identification process, the black pixel dissimilarity of each plane binary composite image is calculated, and the plane that is the basis of the plane binary composite image having the maximum dissimilarity is selected as the dropout plane. Processing,
A dropout color identification processing program which is executed by a computer.

(Supplementary Note 5) In the processing method for identifying the dropout color used in the form dropout process,
A plane acquisition process for acquiring R planes, G planes, and B planes by separating and generating plane images of each RGB system from the form image data;
A threshold value calculation process for obtaining density characteristics for each of the R plane, G plane, and B plane, and automatically calculating a discrimination threshold value for each plane from the density characteristics;
A dropout color identification processing method comprising: a dropout plane selection processing step of selecting a dropout plane from each of the planes using the determination threshold value.

(Supplementary Note 6) In the dropout color identification processing method for identifying the dropout color used in the form dropout process,
A plane acquisition process for acquiring R planes, G planes, and B planes by separating and generating plane images of each RGB system from the form image data;
A luminance signal binarization process for generating a luminance signal image from the form image and binarizing to generate a luminance signal binary image;
Plane 2 for obtaining each of the dynamic determination threshold values acquired in the plane acquisition process, and binarizing each plane based on the determination threshold value to generate each plane binarized image Valuation process,
A combining process for generating a binary combined image of each plane by combining the luminance signal binarized image and each of the plane binarized images for each pixel;
Calculating a pixel frequency of each plane binary composite image, and using the pixel frequency to select a dropout plane from the plane that is the basis of each plane binary composite image; A dropout color identification processing method characterized by comprising:

(Supplementary note 7) A dropout processing device that performs a dropout process using a dropout color identified from a form,
Plane acquisition processing means for separately generating and generating R, G, and B planes of RGB system plane images from the form image data;
Threshold value calculation processing means for obtaining density characteristics for each of the R plane, G plane, and B plane, and automatically calculating a discrimination threshold value for each plane from the density characteristics;
Dropout plane selection processing means for selecting a dropout plane from each of the planes using the determination threshold value.

(Supplementary Note 8) A dropout processing device that performs dropout processing with a dropout color identified from a form,
Plane acquisition processing means for separately generating and generating R, G, and B planes of RGB system plane images from the form image data;
Luminance signal binarization processing means for generating a luminance signal image from the form image and further binarizing to generate a luminance signal binary image;
Plane 2 for obtaining each of the dynamic determination threshold values acquired in the plane acquisition process, and binarizing each plane based on the determination threshold value to generate each plane binarized image Value processing means;
Combining processing means for generating a binary combined image of each plane by combining the luminance signal binarized image and each of the plane binarized images for each pixel;
Drop-out plane selection processing means for calculating a pixel frequency of each of the plane binary composite images and selecting a drop-out plane from the plane that is the basis of each of the plane binary composite images using the pixel frequency; A dropout processing device comprising: a dropout processing device;

(Supplementary Note 9) A dropout processing apparatus that performs dropout processing with a dropout color identified from a form,
Plane acquisition processing means for separately generating and generating R, G, and B planes of RGB system plane images from the form image data;
Luminance signal binarization processing means for generating a luminance signal image from the form image and further binarizing to generate a luminance signal binary image;
Plane 2 for obtaining each of the dynamic determination threshold values acquired in the plane acquisition process, and binarizing each plane based on the determination threshold value to generate each plane binarized image Value processing means;
First combining processing means for combining the luminance signal binarized image and one plane binary image of each plane binarized image for each pixel to generate a first binary combined image; ,
The first binary composite image and one plane binary image that is not processed by the first composite processing means among the plane binary images are combined for each pixel to generate a second binary image. A dropout processing apparatus comprising: a second combining processing unit that generates a combined image.

It is a figure which shows the structural example of the apparatus concerning 1st Embodiment concerning this invention. It is a figure which shows a part of image of the form 2 in 1st Embodiment. It is a figure which shows each histogram example of R plane, G plane, and B plane. It is a figure which shows the example of a dropout plane binary image. It is a figure which shows the example of a 7x7 filter. It is a figure for demonstrating the binarization process of Y signal image, and a figure which shows the example of a Y signal binary image. It is a figure which shows the example of a YR binary image. It is a figure which shows the example of a result by the presence or absence of a blurring correction process. It is a figure for demonstrating the blur correction process with respect to a local binarization process. It is a figure which shows the example of a result by the presence or absence of a blurring correction process. It is a figure for demonstrating the blur correction process with respect to the binarization process by an automatic discrimination threshold method. It is a figure which shows the example of a result by the presence or absence of a blurring correction process. It is a figure which shows the example of a processing flow of the selection process of the dropout plane in 1st Embodiment. It is a figure which shows the structural example of the apparatus concerning this invention in 2nd Embodiment. It is a figure which shows the example of the binary image synthesize | combined by AND operation processing. It is a figure which shows the example of a processing flow of the selection process of the dropout plane in 2nd Embodiment. It is a figure which shows the structural example of the apparatus concerning this invention in 3rd Embodiment. It is a figure which shows the example of the binary image synthesize | combined by the XOR operation process. It is a figure which shows the example of a processing flow of the selection process of the dropout plane in 3rd Embodiment. It is a figure which shows a part of image of the form 2 in 4th Embodiment. It is a figure which shows the structural example of the apparatus concerning this invention in 4th Embodiment. It is a figure for demonstrating the synthetic | combination process of a several AND operation.

Explanation of symbols

2 Form 3, 4 Image image 100, 200, 300, 400 Dropout processing device 101 Image acquisition unit 102 Shadow correction processing unit 103 RGB plane division unit 104 Image storage unit 105 Histogram calculation unit 106 Determination threshold value calculation unit 107 Dropout Plane selection unit 108 Dropout plane binarization processing unit 109 Luminance signal generation unit 110 Luminance signal binarization processing unit 111 Binary image synthesis processing unit 112, 212, 306, 404 Scratch correction processing unit 201 R plane threshold Calculation unit 202 G plane threshold value calculation unit 203 B plane threshold value calculation unit 204 R plane binarization processing unit 205 G plane binarization processing unit 206 B plane binarization processing unit 207, 208, 209 AND operation unit 210, 304, black pixel frequency calculation unit 211,305 Binary image composition processing unit 301,302,303 XOR operation unit 401 YR composition operation unit 402 YG composition operation unit 403 YB composition operation unit

Claims (5)

A program that causes a computer to execute a process of identifying a dropout color used in a form dropout process,
A plain image obtained by removing each RGB system color component from the form image data, the R plane from which the R system color component has been removed, the G plane from which the G system color component has been removed, and the B system color component. A plane acquisition process for generating a B plane;
A luminance signal binarization process for generating a luminance signal image from the form image data and binarizing the luminance signal image to generate a luminance signal binary image;
A threshold value calculation process for obtaining a density characteristic for each plane of the R plane, the G plane, and the B plane, and calculating a discrimination threshold value based on a predetermined formula from the density characteristic of each plane;
A plane binarization process for binarizing each of the R plane, the G plane, and the B plane based on the determination threshold, and generating a binarized image of each plane;
For the luminance signal binarized image and each binarized image of each plane, a corresponding pixel is combined, and a combining process for generating a binary combined image of each plane;
In each binary composite image of each plane, a pixel frequency is calculated, and based on the pixel frequency, a binary composite image showing a tendency of different pixel frequencies is selected from the plane, and the selected 2 Dropout plane selection processing for specifying a color system removed by a plane corresponding to a value composite image as a dropout color;
A dropout color identification processing program which is executed by a computer. In the dropout color identification processing program according to claim 1,
In the synthesis process, an AND operation is performed for each pixel in each of the luminance signal binarized image and the binarized image of each plane to generate a binary synthesized image of each plane.
In the dropout plane selection process, the black pixel frequency of the binary composite image of each plane is calculated, and the color system removed by the plane that is the basis of the binary composite image having the minimum black pixel frequency is calculated. The process identified as a dropout color
A dropout color identification processing program which is executed by a computer. In the dropout color identification processing program according to claim 1,
In the synthesis process, a process of performing an XOR operation for each pixel in the luminance signal binarized image and the binarized image of each plane to generate a binary synthesized image of each plane,
In the drop-out plane selection process, the calculated remaining number how often the black pixels of the binary composite image for each plane, the remaining number how often is removed by plane as the original binary composite image which is the maximum The process of identifying the selected color system as a dropout color,
A dropout color identification processing program which is executed by a computer. A dropout color identification processing method for identifying a dropout color to be used in a form dropout process by a computer,
A plain image obtained by removing each RGB system color component from the form image data, the R plane from which the R system color component has been removed, the G plane from which the G system color component has been removed, and the B system color component. A plane acquisition process for generating a B plane;
A luminance signal binarization process for generating a luminance signal image from the form image data and binarizing the luminance signal image to generate a luminance signal binary image;
A threshold value calculation process for obtaining density characteristics for each of the R plane, G plane, and B plane, and calculating a discrimination threshold from the density characteristics of each plane based on a predetermined formula;
A plane binarization process for binarizing each of the R plane, G plane, and B plane based on the determination threshold, and generating a binarized image of each plane;
For the luminance signal binarized image and each of the binarized images of each plane, a corresponding pixel is combined to generate a binary combined image of each plane;
In each binary composite image of each plane, a pixel frequency is calculated, and based on the pixel frequency, a binary composite image showing a tendency of different pixel frequencies is selected from the plane, and the selected 2 A dropout color identification processing method comprising: a dropout plane selection processing step of specifying a color system removed by a plane corresponding to a value composite image as a dropout color. A dropout color identification processing device for executing a process for identifying a dropout color used in a form dropout process,
A plain image obtained by removing each RGB system color component from the form image data, the R plane from which the R system color component has been removed, the G plane from which the G system color component has been removed, and the B system color component. Plane acquisition processing means for generating a B plane;
Luminance signal binarization processing means for generating a luminance signal image from the form image data and binarizing the luminance signal image to generate a luminance signal binary image;
Threshold calculation processing means for calculating density characteristics for each of the R plane, G plane, and B plane, and calculating a discrimination threshold value based on a predetermined formula from the density characteristics of each plane;
Plane binarization processing means for binarizing each of the R plane, G plane, and B plane based on the determination threshold, and generating a binarized image of each plane;
Synthesis processing means for synthesizing corresponding pixels of the luminance signal binarized image and the binarized image of each plane to generate a binary synthesized image of each plane;
In each binary composite image of each plane, a pixel frequency is calculated, and based on the pixel frequency, a binary composite image showing a tendency of different pixel frequencies is selected from the plane, and the selected 2 A dropout color identification processing device comprising: dropout plane selection processing means for specifying, as a dropout color, a color system removed by a plane corresponding to a value composite image.

Images