JP2001118031A - Optical character reader - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately detect whether or not a read object with a pattern has a character. SOLUTION: A color process part 20 discriminates between chromatic color pixels and achromatic color pixels of an image of a read object 5 obtained by a color image read part 10. A histogram circuit 40 generates histograms of the luminance of the chromatic and achromatic color pixels by every specific number of specific areas of the read object 5, and a histogram analysis part 50 finds binarization slice values for the chromatic color pixels and binarization slice values for the achromatic color pixels for every specified area. A property decision circuit 60 and a filter circuit 70 binarize the image by using the binarization slice values and a print line detecting circuit 90 finds whether or not three is a character for every specified area.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical character reader for optically acquiring an image of a reading medium on which a copy-forgery-inhibited pattern is printed, and dropping out the copy-forgery-inhibited pattern from the image to detect or recognize the presence or absence of a character. It concerns the device.

[0002]

2. Description of the Related Art For example, in a passbook or the like, a copy-forgery-inhibited pattern is printed using various colors. Since these copy-forgery-inhibited patterns are unnecessary for reading characters, it is desirable that they can be seen by humans but cannot be read by a character reading device or the like. For this reason, in a conventional optical character reading device that reads a passbook or the like, a color is used to reflect infrared light, the passbook is irradiated with infrared light, and the reflected infrared light is acquired by an infrared light image sensor. By doing so, the tint block is dropped out. Characters are cut out from the image in which the copy-forgery-inhibited pattern is dropped out, and the presence or absence of characters and character recognition are performed.

[0003]

However, the conventional optical character reader using infrared rays has the following problems. Characters written with a ballpoint pen and characters printed with dye-based ink reflect infrared light, so that when characters are dropped out of the tint block, the characters are also dropped out at the same time. For this reason, there is a problem that a ball-in or a dye-based ink cannot be used for a reading object having a copy-forgery-inhibited pattern. In order to overcome such a problem, the present applicant discloses two types of optical character reading devices in Japanese Patent Application No. 11-173612 filed earlier. Hereinafter, these optical character reading devices are referred to as the first optical character reading device (1) and the second optical character reading device (2), respectively, and the outline will be described below.

(1) Overview of the First Optical Character Reading Device The first optical character reading device optically acquires an image to be read, scans the image of the reading object, and reads the image. Red (hereinafter, referred to as R) and green (hereinafter, referred to as G) representing each target pixel.
Color image reading means for outputting primary color component signals of blue and blue (hereinafter referred to as B) for each pixel;
, And outputs a color identification signal as a result of the identification, based on the primary color component signals of B and B, and outputs the color identification signal of the R, G, and B primary color component signals. A color processing unit for outputting a high signal as a multi-valued luminance signal, a histogram acquisition unit as described below, a histogram analysis unit, and a binarization unit are provided.

[0005] The histogram acquisition means receives the color identification signal and the luminance signal, counts the number of pixels that fall into each gradation level of the luminance signal for each of the chromatic color pixel and the achromatic color pixel, and generates a chromatic color histogram and an achromatic color histogram. It is a means to ask. The histogram analysis unit analyzes the unevenness of the chromatic color histogram and the achromatic color histogram, and determines a first binarized slice value for distinguishing a tint block and a character with respect to a pixel determined as a chromatic color pixel; The second for distinguishing the ground tint and the character with respect to the determined pixel.
This is a means for obtaining a binarized slice value. The binarizing means receives a color identification signal and a luminance signal, and binarizes a pixel whose color identification signal indicates a chromatic pixel into white or black using a first binarization slice value. Are binarized to white or black with the second binarized slice value for pixels indicating achromatic pixels.

(2) Outline of the Second Optical Character Reading Device The second optical character reading device is different from the first optical character reading device in that the binarizing means of the first optical character reading device is replaced by The following pixel attribute determination means and filter means are provided. The pixel attribute determining unit receives the color identification signal, the luminance signal, the first binarized slice value, and the second binarized slice value, and determines whether the target pixel belongs to a pixel of white, black, or another color. Are sequentially determined, and this determination result is output as a pixel attribute. The filter means receives the pixel attribute from the pixel attribute determination means and determines whether the pixel of interest is a pixel forming part of a character or another part based on the pixel attribute of the pixel of interest or the relationship between the pixel of interest and the pixel attributes of surrounding pixels. Are determined, and binarization is performed in which pixels forming part of a character are set to black and pixels forming other parts are set to white. In the first and second optical character reading apparatuses having the above-described configuration, when there is a uniform copy-forgery-inhibited pattern on the entire surface of the reading target, the copy-forgery-inhibited pattern can be dropped out without using infrared rays. However, a passbook or the like may not always have a uniform pattern.

FIG. 2 is a diagram showing an example of a passbook format. An ordinary passbook has a print area 1 in which a date field 1a, a balance field 1b, and the like form a line, and a page mark area 2 in which a plurality of print areas 1 are set and page marks are printed. The color and density of the copy-forgery-inhibited pattern may be different between the page mark area 2 and the print area 1, the color and density of the copy-forgery-inhibited pattern may be different for each line in the print area 1, or the color of the ruled line may be different from the color of the copy-forgery-inhibited pattern. . In such a case, there has been a problem that the line of the page mark area 2 or the print area 1 cannot be accurately detected, or the accuracy of character recognition is deteriorated.

[0008]

According to a first aspect of the present invention, a tint block is printed in color on a surface and characters are written in achromatic color over the tint block. An optical character reading device that optically acquires an image from a printed reading target and detects or recognizes the presence or absence of the character includes a reading unit, an area setting unit, an analyzing unit, and a post-processing unit as described below. I have. The reading means,
A means for optically acquiring an image from the reading target, scanning the image, and outputting a color component signal representing each pixel of the reading target for each pixel. The area setting means sets a range corresponding to an arbitrary number of predetermined areas in the reading target for the image. The analysis means is
It is determined whether the pixel is a chromatic pixel or an achromatic pixel based on the color component signal, and for each of the set ranges, by analyzing the appearance frequency of each pixel, each of the ranges is determined to be the chromatic pixel. A first binarized slice value for distinguishing the tint block and the character with respect to the determined pixel, and a second binary slice value for distinguishing the tint block and the character with respect to the pixel determined as the achromatic pixel. And a value slice value. The post-processing means binarizes each of the pixels using the first and second binarized slice values, and determines whether or not the character exists in each of the predetermined regions based on an image obtained as a result of the binarization. It is means for detecting or recognizing.

According to a second aspect of the present invention, in the optical character reading apparatus, the following color image reading means, color processing means, multivalued image memory, area setting means, histogram acquisition means, histogram analysis means, pixel attribute determination means are provided. ,
A filter means, a binarized image memory and a character processing means are provided. The color image reading means optically acquires an image to be read, performs main scanning and sub-scanning on the image to be read, and outputs red, green, and blue primary color components representing each pixel of the reading object. This is a means for outputting a signal for each pixel. The color processing means identifies whether the pixel is a chromatic pixel or an achromatic pixel based on the red, green, and blue primary color component signals and outputs a color identification signal indicating the identification result. This is a means for outputting the signal with the highest luminance among the B primary color component signals as a multi-valued luminance signal. The multi-valued image memory stores the color identification signal and the luminance signal. The area setting means is means for setting a range corresponding to an arbitrary number of predetermined areas in the reading target in the sub-scanning direction of the image. The histogram acquisition unit inputs the color identification signal and the luminance signal, and for each of the set ranges,
A means for obtaining a chromatic color histogram and an achromatic color histogram by counting the number of pixels that fall into each gradation level of the luminance signal for the chromatic color pixel and the achromatic color pixel.

[0010] The histogram analysis means analyzes the irregularities of the chromatic histogram and the achromatic histogram,
A first binarized slice value for distinguishing the tint block and the character with respect to the pixel determined to be the chromatic pixel;
Means for obtaining a second binarized slice value for discriminating the copy-forgery-inhibited pattern and the character with respect to the pixel determined as the achromatic pixel for each of the set ranges. The pixel attribute determining unit inputs the color identification signal, the luminance signal, the first binarized slice value, and the second binarized slice value for each of the set ranges, and determines whether the pixel of interest is white. This is a means for sequentially determining whether a pixel belongs to a pixel of black or another color, and outputting the determination result as a pixel attribute. Filter means
For each of the set ranges, the pixel attribute is input from the pixel attribute determination means, and the pixel of interest is determined based on the pixel attribute of the pixel of interest or the relationship between the pixel of interest and the pixel attributes of surrounding pixels. It is determined whether pixels constituting part of the character or pixels constituting another part are binarized, and pixels constituting part of the character are set to black and pixels forming the other part are set to white. This is a means for performing for each set range. The binarized image memory stores the binarized result of the filter means as a binarized image. Character processing means
A means for detecting or recognizing the presence / absence of a character in each of the predetermined regions to be read based on the binarized image stored in the binarized image memory.

According to a third aspect of the present invention, there is provided an optical character reading apparatus, wherein the same color image reading means, color processing means, multi-valued image memory, histogram acquisition means, histogram analysis means, pixel attribute determination means as those in the second invention are provided. Filter means, binarized image memory and character processing means,
And an area setting means for setting a range corresponding to an arbitrary number of predetermined areas in the reading target in the main scanning direction of the image. According to a fourth aspect of the present invention, in the optical character reading apparatus, a color image reading unit, a color processing unit, a multi-valued image memory, a histogram acquisition unit, a histogram analysis unit, a pixel attribute determination unit, and a filter unit are provided. A binary image memory and character processing means; and area setting means for setting a range corresponding to an arbitrary number of predetermined areas in the reading object in the main scanning direction and the sub-scanning direction of the image.

According to a fifth aspect of the present invention, in the optical character reading apparatus, the color image reading means, the color processing means, the multi-valued image memory, the area setting means, the histogram obtaining means and the histogram analysis of the second, third or fourth invention are provided. Means for inputting the color identification signal and the luminance signal in the set range read from the multi-valued image memory, wherein the color identification signal represents the first chromatic pixel. The luminance signal is binarized to white or black with the binarized slice value of, and the luminance signal is whitened with the second binarized slice value for the pixel whose color identification signal indicates the achromatic pixel. Alternatively, there are provided a binarizing unit for binarizing to black, and a binarized image memory for storing the binarized result of the binarizing unit as a binarized image.

According to the first to fifth aspects of the present invention, the optical character reading apparatus is configured as described above, so that the area setting means sets an area corresponding to an arbitrary number of predetermined areas with respect to the image to be read. The first and second binarized slice values are determined by the analysis unit or the histogram analysis unit for each set range. Binarization of each pixel is performed using the first and second binarized slice values, and binarization for each range is performed by the pixel attribute determination unit and the filter unit or the binarization unit. Based on the image obtained as a result of the binarization, the presence or absence of a character in each predetermined region to be read is detected or recognized by a post-processing unit or a character processing unit. Therefore, the above problem can be solved.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS First Embodiment FIG. 1 is a block diagram of an optical character reading apparatus showing a first embodiment of the present invention. This optical character reading device is a device that detects, for example, a passbook on which a copy-forgery-inhibited pattern is printed as the reading target 5 and a line on which the characters of the reading target 5 are written or printed, and controls the entire operation (not shown). A control unit, a color image reading unit 10 that is a color image reading unit that optically acquires an image of the reading target 5, and a color processing unit that is a color processing unit connected to an output side of the color image reading unit 10. 20.

The output side of the color processing section 20 is connected to a multi-valued image memory 30 and a histogram circuit 40 having a built-in area setting means and histogram acquisition means. The output side of the histogram circuit 40 is connected to a histogram analysis unit 50 serving as a histogram analysis unit, and the output side of the histogram analysis unit 50 and the output side of the multi-valued image memory 30 are connected to an attribute determination circuit serving as a pixel attribute determination unit. 60 are connected to the input side. A filter circuit 70 is connected to the output side of the attribute determination circuit 60. An output side of the filter circuit 70 is connected to a binarized image memory 80, and an output side of the binarized image memory 80 is
A print line detection circuit 90 which is both a post-processing unit and a character processing unit is connected. From the color processing unit 20 to the histogram analysis unit 50, the color image reading unit 10
Is determined whether the pixel is a chromatic color pixel or an achromatic color pixel based on the signal given from the first pixel, and by analyzing the appearance frequency of each pixel, the first binarized slice value for the chromatic color pixel and the An attribute determination circuit 60, a filter circuit 7
0, binary image memory 80 and print line detection circuit 90
Constitutes a post-processing means for performing binarization of an image to detect the presence or absence of a character.

FIG. 3 shows a color image reading unit 1 shown in FIG.
FIG. 2 is a configuration diagram illustrating an outline of a zero. Color image reading unit 1
0 is a light source 11 that irradiates the reading target 5 with, for example, white light.
And an optical system 12 such as a lens that converges the reflected light from the reading target 5 and, for example, a CCD (Charge Coupled Device). The reflected light obtained through the optical system 12 is converted into R, G, and B light. A line type image sensor 13 for converting color into an analog signal after being divided into three primary colors, and an analog / digital (hereinafter, referred to as A / D) conversion unit 14 connected to an output side of the image sensor 13. are doing. A / D
The conversion unit 14 includes three amplifiers 14a for amplifying three types of electric signals provided from the image sensor 13, respectively.
And three A / D converters 14b which convert the output signals of these amplifiers 14a to digital signals of, for example, 8-bit 256 gradations and output three primary color component signals for each pixel. Thereafter, the three primary color component signals of the digital signal output from each A / D converter 14b are represented by R,
G and B signals.

FIG. 4 is a block diagram showing the color processing unit 20 in FIG. The color processing unit 20 includes a color identification unit 21
And a luminance signal generation circuit 22. The color identification unit 21 receives the R, G, and B signals, has a color look-up table 21a that identifies whether the pixel of interest is a chromatic pixel or an achromatic pixel, and outputs a color identification signal S21 as a result of identification to a histogram circuit. 40 and the multi-valued image memory 30. The luminance signal generation circuit 22 receives the R, G, and B signals, and outputs an 8-bit luminance signal S2 representing the luminance of the target pixel from the R, G, and B signals.
2 is a histogram circuit 40 and a multi-valued image memory 3
It is a circuit that outputs to 0. The multi-valued image memory 30
Color identification signal S21 provided from color identification section 21
And the luminance signal S22.

FIG. 5 is a configuration diagram showing the histogram circuit 40 in FIG. 1, and FIG. 6 is a diagram showing the contents of the histogram memory 43 in FIG. This histogram circuit 4
0 includes a sub-scanning direction counter register 41 as an area setting unit, an address and area switching circuit 42, a histogram memory 43 as a histogram acquisition unit, a register 44, an adder 45, and a data switching circuit 46;
The color identification signal S21 and the luminance signal S22 are sequentially input from the scanning color processing unit 20, and the number of pixels included in each gradation of the luminance signal S22 is determined for each of the pixels identified as chromatic pixels and the pixels identified as achromatic pixels. It has a function to count and calculate the frequency.

The sub-scanning direction counter register 41 is used to create a histogram of each line of the page mark area 2 or the print area 1 shown in FIG. Arbitrary numbers are prepared for the number of. Each sub-scanning direction counter register 41 has a count value in the sub-scanning direction (the transport direction of the reading target 5) in order to set each range of the page mark area 2 or the printing area 1 with respect to the image of the reading target 5. Are respectively set in advance. The address and area switching circuit 42, based on an address switching signal Sc provided from a control unit (not shown), stores a histogram memory 43 provided from the control unit.
Has the function of switching between the clear address Ac for clearing the color, the color identification signal S21 and the luminance signal S22 and outputting the same to the histogram memory 43, and the signal S41 given from the sub-scanning direction counter register 41.
On the basis of the color identification signal S21 and the luminance signal S22.
The writing area to the histogram memory 43 is switched.

The histogram memory 43 assigns each gradation of the luminance signal S22 of each pixel to an address and stores the frequency (the number of pixels) at each gradation. That is, as shown in FIG. 6, 256 levels of gradation 0 to 255 in achromatic pixels are set to achromatic addresses 0 to 255, and 256 levels of gradation 0 to 255 in chromatic pixels are set to chromatic addresses 256 to 512. I have. When there are a plurality of page mark areas 2 or print areas 1, a plurality of storage areas are set for each of the areas so as to store the frequency similarly to FIG. Reading and writing of the frequency of the histogram memory 43 are performed by a control signal R / W.
A register 44 for temporarily storing the frequency read from the histogram memory 43 is connected to an input / output terminal of the histogram memory 43. The adder 45 is connected to the output side of the register 44. The adder 45 performs +1 on the frequency stored in the register 44, and a data switching circuit 46 is connected to an output side of the adder 45. The data switching circuit 46 selects an output signal of the adder 45 or a value of “0” based on the supplied data switching signal Sd and supplies the selected signal to the input / output terminal of the histogram memory 43.

FIG. 7 shows a histogram analysis unit 50 in FIG.
FIG. The histogram analysis unit 50 analyzes the unevenness of the given histogram to obtain a binarized slice value for distinguishing a tint block and a character, and includes a central processing unit (hereinafter referred to as a CPU) 51 and the CPU 5.
1 has an achromatic color maximum component Pareto memory 52, an achromatic color maximum component Pareto change rate memory 53, a chromatic color maximum component Pareto memory 54, and a chromatic color maximum component Pareto change rate memory 55 connected to the bus B.

FIGS. 8A and 8B are diagrams showing an outline of the attribute determination circuit 60 in FIG. The attribute determination circuit 60
The multi-valued image memory 30 receives the luminance signal S22 and the color identification signal S21 from the multivalued image memory 30, and outputs the second signal from the histogram analysis unit 50. , And a binary slice value for a chromatic character, which is a first binary slice value, and a binary slice MAX value from the control unit. Are input, and the attributes of the pixels are obtained by performing the logical determination shown in FIG. 8B.

FIG. 9 is a configuration diagram of the filter circuit 70 in FIG. 1, and FIG. 10 is a diagram showing a pixel arrangement. The filter circuit 70 includes a filter address counter 71 for inputting a count-up signal Cu, and a filter memory 72 connected to the output side of the filter address counter 71.
And a register 73 for storing a determination signal S60 from the attribute determination circuit 60.

A register 74, a register 75, and a register 76 constituting a shift register are sequentially connected to input / output terminals of the filter memory 72. The output terminal of the register 73 is connected to the register 74 at the first stage. Each of the registers 74 to 75 constituting the shift register has 3 registers.
It is composed of stages and outputs the attributes of three pixels. Two-stage output terminals of the three-stage output terminals of the register 74 are connected to input / output terminals of the filter memory 72, and are connected to write pixel attributes to the filter memory 72. Each of the registers 74, 75, and 76 operates to output the attribute determination result of the target pixel and eight peripheral pixels. That is, as shown in FIG. 10, the target pixel # 5 and its surrounding pixels # 1 to # 4, # 6
To # 9, the register 74 sets the pixels # 3 and #
6 and # 9 are output, and the register 75 outputs the attribute determination results of the pixels # 2, # 5 and # 8.
6 outputs the attribute determination results of the pixels # 1, # 4, and # 7.
These registers 74, 75 and 76
Are all connected to the combinational logic circuit 77. A binary image memory 80 connected to the filter circuit 70 stores a binary image.
The print line detection circuit 90 is a circuit that scans the print line area of the binary image stored in the binarized image memory 80, detects, for example, the presence or absence of a character, and outputs the detection result to the outside.

Next, the operation of the optical character reader of FIG. 1 will be described. FIG. 11 is a diagram illustrating the scanning direction of the reading target 5. In the passbook or the like to be read 5, the copy-forgery-inhibited pattern is printed in various colors, and characters or the like are printed or written. When the reading target 5 is transported and reaches the reading position, the reading target 5 is illuminated by the light source 11, and the light reflected by the reading target 5 is input to the image sensor 13 via the optical system 12. The image sensor 13 reads the image of the reading target 5, scans in the main scanning direction in FIG. 11, and sequentially converts the image into electrical signals of three primary color components of R, G, and B. Three
The electric signal of the primary color component is supplied to the amplifier 1 in the A / D converter 14.
The signal is amplified by the A / D converter 4a and converted into R, G, and B signals of 8-bit 256 gradations by the respective A / D converters 14b. The R, G, and B signals are output to the color processing unit 20.

Color identification unit 21 in color processing unit 20
Inputs the R, G, and B signals, and looks up the look-up table 21 based on a combination of the values of the R, G, and B signals.
Based on a, the target pixel is identified as a chromatic pixel or an achromatic pixel, and a color identification signal S21 of the identification result is output. That is, the difference | VR-V between the value VR of the R signal and the value VG of the G signal.
G | is equal to or less than the value of Δ obtained experimentally, and the value VR of the R signal
Is less than or equal to the value of Δ, and the difference | VB between the value VB of the B signal and the value VG of the G signal.
When −VG | is equal to or smaller than the value of Δ, the target pixel is regarded as an achromatic color, and the color identification signal S21 is set to “L” and output.
When the combination of the values of the R, G, and B signals is other than this,
The pixel of interest is regarded as a chromatic color, and the color identification signal S21 is set to "H" and output. The luminance signal generation circuit 22 in the color processing unit 20 receives the R, G, and B signals, and outputs a signal indicating the highest value among the R, G, and B signals as the luminance signal S22 of the pixel of interest. . Color identification signal S
21 and the luminance signal S2 output by the luminance signal generation circuit 22
2 is stored in the multi-valued image memory 30 and input to the histogram circuit 40.

The histogram circuit 40 includes a color identification unit 2
A histogram of chromatic pixels and achromatic pixels is created using the color identification signal S21 input from Step 1 and the luminance signal S22 input from the luminance signal generation circuit 22. Hereinafter, an operation of creating the histogram will be described. First, the address and area switching circuit 42 is supplied with an address switching signal Sc for selecting an address signal Ac from a control unit (not shown), and the data switching circuit 46 is supplied with a data switching signal Sd for selecting “0”. . Thereafter, a clear address signal Ac is sequentially supplied to the address and area switching circuit 42 from a control unit (not shown), and a control signal R / W designating writing is supplied to the histogram memory 43 in synchronization therewith. Thereby, the histogram memory 43 is cleared.

When the clearing of the histogram memory 43 is completed, the control section supplies an address switching signal Sc for selecting the color identification signal S21 and the luminance signal S22 to the address and area switching circuit 42, and an adder 45 to the data switching circuit 46. Is provided. In this state, the output terminal of the adder 45 is connected to the input / output terminal of the histogram memory 43. On the other hand, in each sub-scanning direction counter register 41, a control counter value which is a range of the page mark area 2 or the row in the sub-scanning direction is set in advance by the control unit. When the pixel of the reading target 5 is included in any of the management counter values, the address and area switching circuit 42 functions to form a histogram in a corresponding area of the histogram memory 43. When the color identification signal S21 and the luminance signal S22 are input from the color processing unit 20, they are given to the address input terminal of the histogram memory 43. At this time, the histogram memory 43 reads the value stored at the address corresponding to the color identification signal S21 and the luminance signal S22 as the frequency based on the control signal R / W instructed to read from the control unit at this time, and Store. The adder 45 adds +1 to the value stored in the register 44 and supplies the value to the input / output terminal of the histogram memory 43 via the data switching circuit 46.

The histogram memory 43 stores an adder 45 based on the control signal R / W given from the controller at this time.
Is written in the address corresponding to the color identification signal S21 and the luminance signal S22. For example, if the target pixel is an achromatic color and its luminance gradation is 150, the contents of the address 150 of the histogram memory 43 in FIG. 6 are read out, 1 is added to the value, and the address 150 is again read. Is stored in When the pixel of interest is a chromatic color and its luminance gradation is 150, the address 406 (= 150 + 25) of the histogram memory 43 in FIG.
6) The content of the address is read, 1 is added to the value, and the value is stored again at the address 406. By performing such processing each time the color identification signal S21 and the luminance signal S22 are obtained by scanning the reading target 5, the frequency of each gradation of the maximum component is accumulated. As shown in FIGS. 12A and 12B, two types of histograms, that is, an achromatic color histogram 91 and a chromatic color histogram 92 with the maximum frequency being 100%, are obtained for each page mark area 2 or each line.

FIGS. 12A and 12B are diagrams showing an achromatic color histogram and a chromatic color histogram. When black characters are printed or described on the reading target 5 having the copy-forgery-inhibited pattern, the frequency of the achromatic color histogram 91 is close to the black position (0 in luminance) and the vicinity of the white position (255 in luminance). concentrate. When a pixel of a character is on a white background, the pixel is processed as an achromatic color.However, when a pixel of a character is printed or described on a copy-forgery-inhibited pattern, the ground color is transmitted and colored. Is processed as a chromatic color though it is an achromatic color. Hereinafter, the characters on the white background are referred to as achromatic characters, and the characters on the copy-forgery-inhibited pattern are referred to as chromatic characters.

FIG. 13 is a diagram showing the processing contents of the histogram analyzer 50. The histogram analysis unit 50, which has input the achromatic color histogram 91 and the chromatic color histogram 92 for each page mark area 2 or each line, analyzes these irregularities, and converts the binary slice value for the chromatic character and the 2 for the achromatic character. Although the quantified slice value is obtained for each page mark area or each line, it is difficult to analyze the achromatic histogram 91 and the chromatic histogram 92 because there are local irregularities. For this reason, the histogram analyzer 50 compares the achromatic color tomogram 91 and the chromatic color histogram 92 with each other in FIG.
Achromatic color Pareto calculation processing and chromatic color Pareto calculation processing (1), achromatic color Pareto change rate calculation processing and chromatic color Pareto change calculation processing (2), achromatic color Pareto change rate analysis processing, and chromatic color Pareto change rate analysis processing Perform (3). The details of each process are described below.

(1) Achromatic Color Pareto Calculation Process and Chromatic Pareto Calculation Process FIG. 14 is a flowchart showing the achromatic color Pareto calculation process. FIG. 15 is a flowchart showing the achromatic color maximum corresponding to FIG. It is a component Pareto diagram. The histogram analysis unit 50 uses the histogram circuit 40 in the CPU 51 controlled by software, and executes the processing in step ST in FIG.
In P1 to step STP6, an achromatic color Pareto calculation process is performed.

First, in step STP1, the CPU
An address and area switching circuit 42 from the control unit 51 via the control unit
, And the address and area switching circuit 42 selects the address signal Ac. Step S
At TP2, the value of the address counter Adr corresponding to the address of the histogram memory 43 is set to 0, and the data stored in the address 0 of the histogram memory 43 (corresponding to the 0 level of the luminance signal) is read. In step STP3, the CPU 51 stores the read data at the same address in the achromatic maximum component pareto memory 52. Step ST
In P4, the CPU 51 increments the value of the address counter by +1. Then, in step STP5, the data at the address Adr of the histogram memory 43 is read out, and (A
d-1) The read data is added to the stored data at the address. The CPU 51 stores the result of this addition at the address Adr of the achromatic maximum component pareto memory 52. Step S
By TP6, the operations of steps STP4 and STP5 are repeated until the address counter Adr becomes "255", and a Pareto representing the cumulative frequency is obtained.

FIG. 16 is a flowchart showing the chromatic color Pareto calculation process, and FIG. 17 is a chromatic color maximum component Pareto diagram expressed as a percentage corresponding to FIG. 12B.
The histogram analysis unit 50 determines in step STP1 of FIG.
From step 1 to step STP16, a chromatic color Pareto calculation process is performed. First, in step STP11, the CPU 5
1 to the address and area switching circuit 42 via the control unit,
Select the address signal Ac. Step STP12
, The value of the address counter Adr corresponding to the address of the histogram memory 43 is set to “256”, and the data stored at the address 256 of the histogram memory 43 is read. In step STP13, the CPU 51 stores the read data at the same address of the chromatic color maximum component Pareto memory 54. In step STP14, the CPU
51 increments the value of the address counter Adr by +1. In step STP15, the data at the address Adr in the histogram memory 43 is read, and the read data is added to the storage data at the address (Ad-1) immediately before the chromatic color maximum component Pareto memory 54. CPU5
1 stores this addition result in a chromatic color maximum component Pareto memory 5
4 is stored in the address Adr. In step STP16, the operations in steps STP14 and STP15 are repeated until the value of the address counter Adr becomes "511", and a Pareto representing the cumulative frequency is obtained.

(2) Achromatic color Pareto change rate calculation processing and chromatic Pareto change rate calculation processing FIG. 18 is a flowchart showing the achromatic color Pareto change rate calculation processing, and FIG. 19 is a percentage expression corresponding to FIG. It is a figure showing the achromatic color maximum component Pareto change rate.
The histogram analysis unit 50 performs the achromatic color Pareto change rate calculation process in steps STP21 to STP29 in FIG. 18 by the CPU 51 controlled by software.

In the achromatic color Pareto change rate calculation processing, the maximum achromatic color component Pareto is divided into, for example, 10 gradations.
Is assumed, and in step STP21, the CPU
51 is 10 in the change rate section variable kukan indicating the section.
Is set. In step STP22, the CPU 5
1 sets “0” to a variable j indicating the address of the achromatic maximum component Pareto change rate memory 53, and sets STP2.
In step 3, "255" is set as a variable i indicating the address of the achromatic maximum component pareto memory 52. Step STP2
In step 4, the CPU 51 checks the value of (i-kukan). If the value of (i-kukan) is not negative, the CPU 51 checks in step STP25 for the maximum achromatic color component Pareto memory 52.
From the value stored at the address i of (i-ku
kan) The value stored at the address is subtracted, and the subtracted result is stored at the address j of the achromatic maximum component Pareto change rate memory 53.

On the other hand, in step STP24, (i-kuk
If the value of an) becomes negative, there is no data stored in the achromatic maximum component pareto memory 52 to be referred to. The value of the address 0 in the memory 52 is subtracted, and the result is stored in the address j of the achromatic maximum component Pareto change rate memory 53. Step STP25
Or, after step STP26, step STP2
In step 7, 1 is added to the variable j, and in step STP28, the value of the change rate section variable kukan is subtracted from the variable i. Step ST
By P29, the operations of steps STP24 to STP28 are repeatedly performed, and the process ends when the variable becomes negative.
As described above, the difference between the maximum values of the cumulative frequencies of adjacent sections in the achromatic pareto is sequentially obtained and stored in the achromatic maximum component pareto change rate memory 53, and the smoothed achromatic maximum as shown in FIG. The component Pareto change rate is determined.

FIG. 20 is a flowchart showing the chromatic color Pareto change rate calculation processing, and FIG. 21 is a diagram showing the chromatic color maximum component Pareto change rate expressed as a percentage corresponding to FIG. The histogram analysis unit 50 is a CPU 51 controlled by software, and performs step S in the case of an achromatic color.
Similar to TP21 to step STP29, the chromatic color rate change rate calculation processing is performed in steps STP31 to STP39 in FIG. 20 to obtain the smoothed chromatic color maximum component Pareto change rate as shown in FIG.

(3) Achromatic Color Pareto Change Rate Analysis Processing and Chromatic Pareto Change Rate Analysis Processing The histogram analyzer 50 scans the achromatic color maximum component Pareto change rate and the chromatic color maximum component Pareto change rate obtained so far. By analyzing the unevenness, both the binary slice value for achromatic characters and the binary slice value for chromatic colors are obtained. First, the chromatic color Pareto change rate for obtaining the binary slice value for chromatic characters The analysis process will be described.

FIG. 22 is a flowchart showing the chromatic color Pareto change rate analysis processing (part 1), FIG. 23 is a flowchart showing the chromatic color Pareto change rate analysis processing (part 2), and FIG. It is a flowchart which shows a coloring Pareto change rate analysis process (3). FIGS. 22 to 24 show the entire chromatic color Pareto change rate analysis processing which is linked by the connector. The CPU 51 of the histogram analysis unit 50 performs steps STP41 to STP68 of FIGS. 22 to 24 and analyzes the unevenness of the chromatic maximum component Pareto change rate of FIG. 21 to calculate a binary slice value for achromatic characters. I do.

First, in step STP41, the CP
U51 sets, as initial settings, a block detection flag indicating that a block has been detected, a decrease flag, an increase flag, and a flat flag indicating that the Pareto change rate is not increased, decreased, or neither, to 0, A decrease counter for counting the number of continuous decreasing sections, an increasing counter for counting the number of continuous increasing sections, a flat counter for counting the number of continuous flat sections, a counter cnt for counting the number of block candidate sections, and a detected block number. The indicated block number is initialized to 0. Step ST
In P42, the maximum value -1 of the section number is set as the section variable i for scanning the chromatic color maximum component change rate. In this embodiment, since the gray level of the luminance signal S22 is 256 and the section width is 10, the number of sections is 25 (0 to 25). Therefore, 24 is set in the section variable i. In step STP43, the CPU 51 determines that the section variable i
The change rate of the section indicated by is checked. At this time, if the block detection flag is 0, the block is not detected, and the change rate is smaller than the slice level Th1 determined by experiments or the like, the data in the section i is excluded as noise in step STP44, The counter cnt, the decrease counter, and the increase counter are initialized, and the process proceeds to the scanning in the next section. Otherwise, in step STP45, the block detection flag is checked.

If the block detection flag is 0 and no block has been detected yet, the CPU 51 proceeds to step S
At TP46, 1 is added to the counter cnt, and step STP
At 47, it is determined whether the block is an effective block by comparing the value of the counter cnt with a block constant determined in advance through experiments, evaluations, and the like. If the value of the counter cnt is larger than the block constant, it is determined that a valid block start point has been detected, and in step STP48, a block detection flag is set to 1, 1 is added to the block No. Lower gradation level value of 25
In the next step STP49, the section attribute is determined as 5-i * kukan. In contrast, step STP45
In the determination of the block detection flag, if the block detection flag is set to 1 and the starting point of the block has already been detected, and if it is determined in step STP47 that the block is not a valid block, the section attribute in step STP49 Jump to the decision.

FIG. 25 is a flowchart showing step STP49 in FIG.
FIG. 5 is a diagram showing a determination condition for attribute determination in FIG. In the section attribute determination in step STP49, the CPU 51 subtracts the chromatic color maximum component Pareto change rate (i + 1) in the section variable (i + 1) from the chromatic color maximum component Pareto change rate (i) in the section variable i, and the value is obtained. If negative, it is determined that the rate of change has decreased, and it is determined that the interval is a decrease section.
At 50, the section attribute flag (decrease flag) is set to decrease, 1 is added to the decrease counter, and the flat counter is reset to 0. If the decrease flag is 1 in the section attribute determination in step STP49, the chromatic color maximum component Pareto change rate (i + 1) in the section variable (i + 1) is subtracted from the chromatic color maximum component Pareto change rate (i) in the section variable i. If the value is a positive value, it is determined that the period has changed from decreasing to increasing,
In step STP51, increment is set in the section attribute flag (increase flag), 1 is added to the increment counter, and the flat counter is reset to 0. Step STP
In the section attribute determination of 49, when the decrease flag is 1, the chromatic color maximum component Pareto change rate (i + 1) in the section variable (i + 1) is subtracted from the chromatic color maximum component Pareto change rate (i) in the section variable i. If the value is smaller than the flat constant obtained in advance through experiments or the like, it is determined that the change rate has hardly changed, and in step STP52, flat is set in the section attribute flag (flat flag), and the flat counter is set. , And reset the flat flag to 0.

In step STP53 of FIG.
The PU 51 checks the block detection flag. When the block detection flag is 0, the PU 51 scans the next section.
After subtracting 1 from the interval variable i, the process jumps to step STP43 via the connector. When the block detection flag is 1 and the start point of the block has already been detected, the increase or decrease of the section before the section specified by the variable section i is checked in steps STP54 to STP58. For example, in step STP54, the value of the decrease counter is compared with, for example, 1 of the value of the decrease parameter determined by an experiment or the like, and if the length of the decrease section (= the value of the decrease counter) is larger than the value of the decrease parameter, Sets the decrease flag to 1 in step STP55 assuming that the decrease is continuing. In step STP56, the value of the increase counter is compared with the value of the increase parameter, eg, 2, determined by an experiment or the like, and the length of the increase section (= the value of the increase counter) is larger than the value of the increase parameter. In this case, it is determined that the increase is continuing, and the increase flag is set to 1 in step STP57. Further, in step STP18,
The value of the flat counter is compared with the value of the flat parameter determined by experiment or the like, for example, 2. If the flat section (= flat counter value) is larger than the flat parameter, it is determined that flat is continued, and step At STP59, 1 is set to the flat flag.

At step STP60 after steps STP54 to STP58 and step STP59, C
The PU 51 determines whether there is a valley or a flat portion using the valley detection condition. That is, the case where the decrease flag is 1 and the increase flag is 1, that is, the chromatic color maximum component Pareto change rate changes from decrease to increase, and the case where the flat flag is 1, that is, the chromatic color maximum component Pareto change rate decreases from flat to flat. In the case of turning over and the case where the section variable i becomes 0 and there is no section to be scanned, it is determined that the condition satisfying the valley is satisfied and the block is detected. Then, the CPU 51 proceeds to step ST.
In P61, the upper gradation value level of the block is set to “25”.
5- (i + 3) * kukan ". On the other hand, if the condition for satisfying the valley is not satisfied, it is determined that a block has not been detected, and in step STP62, the same value as the lower gradation level value is stored in the upper gradation level value of the block. I do.

In step STP63, the upper gradation level value and the lower gradation level value of the block are compared, and when both the level values match, the block is regarded as still in the middle and the section variable After subtracting 1 from i, the process returns to step STP43 to continue scanning in the next section. On the other hand, if the levels do not match, it is determined that a block has been found, and the operation of detecting the block is terminated, and in step STP64, a block candidate, a decrease flag, an increase flag, a flat flag, a decrease counter, an increase counter, Each value of the flat counter is initialized to zero. Step ST after completion of the initialization of step STP64
In P65, the value of the interval variable i is checked, and
Is not 0 and there is an unscanned section, in step STP66, a new block detection is performed by starting a new scan at a position returned by a predetermined section. The return section is determined by an experiment or the like, and is, for example, three sections. If the section variable i is 0, the scanning ends.

At step STP67 after the scanning is completed, a gradation level at which the value of the maximum chromatic color component Pareto is, for example, 50% is calculated, and for example, 30 of the slide parameter determined by experiment or the like is subtracted from the calculation result. The value obtained is the gradation level of the tint block. In the chromatic color maximum component pareto of FIG. 17, the gradation level at which the cumulative frequency is 50% is “188”, and the value obtained by subtracting the slide parameter from that is “158” is the copy-forgery-inhibited pattern gradation level. In the next step STP68 following the step STP67, of the upper gradation levels of the detected block,
A block having a value closest to the tint block gradation level is regarded as a chromatic character block, and the upper gradation level of the block is set as a binarized slice value. In the example of the chromatic color maximum component Pareto change rate shown in FIG. 21, three blocks B1, B
2 and B3, the upper gradation level of the block B1 closest to the copy-forgery-inhibited pattern gradation level “158” is “12”.
5 ", and this" 125 "is the binary slice value for chromatic characters. The binarized slice value for a chromatic character is sent to the attribute determination circuit 60.

FIG. 26 is a flowchart showing the achromatic color Pareto change rate analysis processing (part 1), FIG. 27 is a flowchart showing the achromatic color Pareto change rate analysis processing (part 2), and FIG. It is a flowchart which shows a coloring Pareto change rate analysis process (3). FIGS. 26 to 28 show the entire achromatic color Pareto change rate analysis process linked by the connector. The CPU 51 of the histogram analysis unit 50 performs steps STP71 to STP97 of FIGS. 26 to 28 to analyze the unevenness of the chromatic maximum component pareto change rate of FIG. 21 to calculate a binary slice value for achromatic characters. I do.

Steps STP71 to STP96 of steps STP71 to STP97 perform the same processing as steps STP41 to STP66 in the chromatic color Pareto change rate analysis processing except that the chromatic color changes to an achromatic color.
When the scanning in steps STP41 to STP66 is completed, in step STP97, the block in which the upper gradation level of each detected block is closest to, for example, “200” of the MAX value of the preset binarized slice, The block is regarded as an achromatic character block, and the upper gradation level of the block is set as a binary slice value for the achromatic character. In the example of FIG. 19, three blocks B4, B5, and B6 are detected in the achromatic color maximum component Pareto change rate, but the upper gradation level “19” closest to the slice MAX value “200” is detected.
The block B5 having "5" is regarded as an achromatic character, and the gradation level "195" is obtained as a binary slice value for the achromatic character. The binarized slice value for the achromatic character is sent to the attribute determination circuit 60. The above is the operation of the histogram analysis unit 50.

The histogram analysis unit 50 outputs a binary slice value for a chromatic character and a binary slice value for an achromatic character, and outputs a luminance signal S22 and a color identification signal for each pixel from the multi-valued image memory 30. S21 is output. The attribute determination circuit 60 converts the binary slice value for chromatic character, the binary slice value for achromatic character, the luminance signal S22, the color identification signal S21, and the binary slice MAX value given from the control unit. By comparison, the attribute of the target pixel # 5 is determined to be three types of attributes of white, black or color. For example, attention pixel # 5
When the color identification signal S21 of “1” is “0” and indicates an achromatic color,
The luminance signal S22 is compared with the achromatic character binarized slice value. If the luminance signal S22 is larger than the achromatic character binarized slice value, “11” is set as the pixel attribute.
The determination result S60 of (white) is output. In the opposite case, a determination result S60 of “00” (black) is output as the pixel attribute.

The color identification signal S2 of the target pixel # 5
When 1 is "1" and indicates a chromatic color, the luminance signal S22 is compared with the binarized slice MAX value, and the luminance signal S22 is compared with the luminance signal S22.
Is higher, the determination result S60 of “11” (white) is output as the pixel attribute. Conversely, if the luminance signal S22 is smaller than the binary slice MAX value and larger than the chromatic character binary slice value, the determination result S60 of "01" (color) is output as the pixel attribute. I do. If the luminance signal S22 is smaller than the binary slice value for chromatic character, a determination result S60 of "00" (black) is output as the pixel attribute. Conversely, if the luminance signal S22 is smaller, the pixel attribute is "00". The determination result S60 of “(black) is output. Determination result S60 indicating the attribute of the pixel output from attribute determination circuit 60
Are sequentially supplied to the filter circuit 70 and are taken in by a shift operation. Register 74 in filter circuit 70-
76 is a pixel of interest # 5 and its surrounding pixels # 1 to # 4, #
The attributes of 6 to # 9 are shown in the combinational logic circuit 77.

FIG. 29 is a diagram showing the logic of the combinational logic circuit 77. The combinational logic circuit 77 determines whether the attribute of the target pixel # 5 is “11” (white) or “00”.
In the case of (black), white or black according to the attribute is output as a filter output regardless of the peripheral pixels.
On the other hand, when the attribute of the target pixel # 5 is “01” (color), the total number of black pixels of the surrounding pixels is compared with, for example, three filter parameters determined by experiments or the like.
If it is larger than this, it is considered that a part of the character is erroneously determined to be a color, and the target pixel # 5 is converted from color to black and output. Conversely, when the number of black pixels is smaller than 3, the target pixel # 5 is converted from color to white and output. The filter output output from the combinational logic circuit 77 is written to the binary image memory 80 as a binary image to be read.
The print line detection circuit 90 is provided in the binarized image memory 80.
The digitized image is scanned to detect whether there is a character printed or described on a predetermined line.

As described above, in the first embodiment,
A color image acquisition unit 10, a color processing unit 20, a histogram circuit 40, and a histogram analysis unit 50 are provided. The histogram circuit 40 is provided with a sub-scanning direction counter register 41, and the image of the reading target 5 is , The pixels are divided into achromatic and chromatic colors, histograms for the achromatic and chromatic colors are created in the page mark area 2 and an arbitrary number of areas for each row, and the histograms are analyzed for And a binarized slice for achromatic color are used. Therefore, even when the color and density of the copy-forgery-inhibited pattern for each page mark area 2 and each line are different, and when the color of the ruled line is different from the color of the copy-forgery-inhibited pattern, characters are reliably left for each page mark area 2 and each line. You can drop out only the tint block.

Second Embodiment FIG. 30 is a configuration diagram of an optical character reading apparatus showing a second embodiment of the present invention, and FIG. 31 is a configuration diagram showing a histogram circuit 40A in FIG. 1 and 5 showing the first embodiment are denoted by common elements. This optical character reading device is obtained by changing the histogram circuit 40 of the optical character reading device of the first embodiment to a histogram circuit 40A, and the other configuration is the same as that of FIG. Histogram circuit 40A
As shown in FIG. 31, an arbitrary number of main scanning direction counter registers 47 are provided in place of the sub-scanning direction counter register 41 as shown in FIG. 31, and the other configuration is the same as that of FIG. The main scanning direction counter register 47 is a register for setting a counter value in the main scanning direction (pixel position of the sensor 13) for controlling a writing area to the histogram memory 43.

Next, the operation of the optical character reader of FIG. 30 will be described. The color image reading unit 10 and the color processing unit 20 perform the same operations as in the first embodiment. In the histogram circuit 40A, first, an address and area switching circuit 42 is supplied with an address switching signal Sc for selecting an address signal Ac from a control unit (not shown), and a data switching signal for selecting "0" is supplied to a data switching circuit 46. Sd is provided. Thereafter, a clear address signal Ac is sequentially supplied to the address and area switching circuit 42 from a control unit (not shown), and a control signal R / W designating writing is supplied to the histogram memory 43 in synchronization therewith.
Thereby, the histogram memory 43 is cleared.

When the clearing of the histogram memory 43 is completed, the control unit supplies an address switching signal Sc for selecting the color identification signal S21 and the luminance signal S22 to the address and area switching circuit 42, and an adder 45 to the data switching circuit 46. Is provided. In this state, the output terminal of the adder 45 is connected to the input / output terminal of the histogram memory 43. On the other hand, in the main scanning direction counter register 47, a control counter value that is in the range in the main scanning direction is set in advance by the control unit. When an image of the reading target 5 is obtained, a color identification signal S21 and a luminance signal S22 for each pixel are supplied to an address input terminal of the histogram memory 43. If the position of the pixel falls within the range set in a certain main scanning direction counter register 47, the frequency of the histogram is accumulated in the storage area of the histogram memory 43 corresponding to the main scanning direction counter register 47. In other words, a histogram is created for each of the arbitrary number of main scanning direction counter registers 47, and pixels at locations not set in the main scanning direction counter register 47 are excluded from the histogram. The register 44, the adder 45, and the data switching circuit 46
3 operates in the same manner as in the first embodiment, and generates the same number of histograms as the main scanning direction counter register 47.
The multivalued image memory 30, the histogram analyzer 50, the attribute determination circuit 60, the filter circuit 70, the binarized image memory 80, and the print line detection circuit 90 operate in the same manner as in the first embodiment.

As described above, in the second embodiment,
Since the main scanning direction counter register 47 is provided instead of the sub scanning direction counter register 41, a histogram is created, for example, when the reading target 5 has a date column 1a and a balance column 1b as shown in FIG. The area can be limited to the date column 1a or the balance column 1b, and it is possible to drop out only the copy-forgery-inhibited pattern while leaving the characters more accurately than the second optical character reading device, and to detect an appropriate print line. Will be possible. In addition, the processing speed can be improved by limiting the area where the histogram is created.

Third Embodiment FIG. 32 is a block diagram showing an optical character reading apparatus according to a third embodiment of the present invention, and FIG. 33 is a block diagram showing a histogram circuit 40B shown in FIG. 1 and 5 showing the first embodiment are denoted by common elements. This optical character reading device is obtained by changing the histogram circuit 40 of the optical character reading device of the first embodiment to a histogram circuit 40B, and the other configuration is the same as that of FIG.

As shown in FIG. 33, the histogram circuit 40B includes an arbitrary number of main scanning direction counter registers 48a and a sub scanning direction counter register 48b paired with the main scanning direction counter register 48a. It is provided in place of 41. Other configurations are the same as those in FIG. Main scanning direction counter register 48a
Each pair of the sub-scanning direction counter register 48b is a register for setting a counter value for controlling the writing area to the histogram memory 43.

Next, the operation of the optical character reader of FIG. 32 will be described. The color image reading unit 10 and the color processing unit 20 perform the same operations as in the first embodiment. In the histogram circuit 40B, first, the address and area switching circuit 42 is supplied with an address switching signal Sc for selecting the address signal Ac from a control unit (not shown), and the data switching circuit 46 is supplied with a data switching signal for selecting "0". Sd is provided. Thereafter, a clear address signal Ac is sequentially supplied to the address and area switching circuit 42 from a control unit (not shown), and a control signal R / W designating writing is supplied to the histogram memory 43 in synchronization therewith.
Thereby, the histogram memory 43 is cleared.

When the clearing of the histogram memory 43 is completed, the control unit supplies the address and area switching circuit 42 with the address switching signal Sc for selecting the color identification signal S21 and the luminance signal S22, and the data switching circuit 46 with the adder 45. Is provided. In this state, the output terminal of the adder 45 is connected to the input / output terminal of the histogram memory 43. On the other hand, in the main scanning direction counter register 48a, a management counter value which is a range in the main scanning direction is set in advance by the control unit, and in the sub scanning direction counter register 48b, the range in the sub scanning direction is set by the control unit. Is set. When acquiring the image of the reading target 5, the color identification signal S21 and the luminance signal S22 for each pixel are input to the address input terminal of the histogram memory 43,
Given. If the position of the pixel falls within the range set in a certain main scanning direction counter register 48a and a certain sub scanning direction counter register 48b, the main scanning direction counter register 48a and the sub scanning direction counter register 48a
The frequency of the histogram is accumulated in the storage area of the histogram memory 43 corresponding to b. That is, a histogram is created for each of the arbitrary number of the main scanning direction counter registers 48a and the sub-scanning direction counter registers 48b, and pixels in places other than those are excluded from the histogram. The register 44, the adder 45, and the data switching circuit 46 operate on the histogram memory 43 in the same manner as in the first embodiment, and generate the same number of histograms as the main scanning direction counter register 47. Multi-valued image memory 30, histogram analyzer 50, attribute determination circuit 6
0, the filter circuit 70, the binarized image memory 80, and the print line detection circuit 90 operate in the same manner as in the first embodiment.

As described above, in the third embodiment,
The main scanning direction counter register 48a and the sub-scanning direction counter register 48b are used to set a rectangular area with respect to the reading target 5. Therefore, in the page mark area 2 and each line, an arbitrary area such as the date section 1a and the balance section 1b is set. Can be created, the characters can be left more accurately, and only the copy-forgery-inhibited pattern can be dropped out.

Note that the present invention is not limited to the above embodiment, and various modifications are possible. For example, there are the following modifications. (I) In the first to third embodiments, the print line detection circuit 90
Is used to detect the presence or absence of a character. However, for example, a character recognition device can be provided by providing a character cutout unit or a character recognition unit.

(Ii) FIG. 34 is a configuration diagram of an optical character recognition device according to a modification of the first embodiment, and FIG. 35 is a configuration diagram showing the binarization circuit 100 in FIG. In the first embodiment, the attribute determination circuit 60 and the filter circuit 70
However, instead of the attribute determination circuit 60 and the filter circuit 70, a binarization circuit 10 as shown in FIG.
It can also be configured using 0. Binarization circuit 100
Is the luminance signal S2 stored in the multi-valued image memory 30.
2 is a circuit for converting into a binary image composed of white or black using the binarized slice value given from the histogram analysis unit 50.
Comparator 102, achromatic character register 103, comparator 1
04 and a switching circuit 105. Comparator 102
Is the luminance signal S read from the multi-level image memory 30.
22 is compared with the data S101 given from the register 101, and the output side is connected to the switching circuit 105. The comparator 104 is a multi-valued image memory 3
A comparison is made between the luminance signal S22 from 0 and the data S103 given from the register 103, and the output side is connected to the switching circuit 105. The switching circuit 105
When the color identification signal S21 read from the multi-valued image memory 30 indicates a chromatic pixel, the output signal of the comparator 102 is selected. When the color identification signal S21 indicates an achromatic pixel, the comparator 104 selects the output signal.
Is selected and output.

The binarizing circuit 100 operates as follows. The chromatic character register 101 stores the binary slice value S101 for the chromatic character,
02, the comparator 102 compares the luminance signal S22 with the binarized slice value S101 to generate a binary signal S102 indicating white or black. The achromatic character register 103 stores the binary slice value S103 for an achromatic character and supplies the binary slice value S103 to the comparator 104, which compares the luminance signal S22 with the binary slice value S103. , A binary signal S104 indicating white or black is generated. The switching circuit 105 selects the binary signal S102 when the color identification signal S21 indicates the chromatic color of the pixel, and selects the binary signal S104 when the color identification signal S21 indicates the achromatic color of the pixel. It is output as a value image signal S100. Even if the binarizing circuit 100 is used, the binarized image is stored in the binarized image memory 80,
The same effects as in the first embodiment can be obtained. Furthermore, the second
Also, the third embodiment can be configured using a binarization circuit 100 instead of the attribute determination circuit 60 and the filter circuit 70.

[0066]

As described above in detail, according to the first aspect, reading means for obtaining an image to be read,
Area setting means for setting a range corresponding to an arbitrary number of predetermined areas in the reading target to an image, analysis means for obtaining first and second binarized slice values for each of the ranges, and post-processing means Since the type character reading device is configured, processing for detecting or recognizing the presence or absence of a character is limited to only a predetermined area.
And the second binarized slice value can be accurately obtained, the accuracy of character presence detection and recognition is improved, and the processing speed is increased.

According to the second to fourth inventions, a color image reading means, a color processing means, a multi-valued image memory, an area setting means, a histogram acquisition means, a histogram analysis means, a pixel attribute determination means, a filter means, a binary According to a fifth aspect of the present invention, a color image reading device, a color processing device, a multi-valued image memory, an area setting device, a histogram obtaining device, and a histogram analyzing device are provided. Since the binarizing means, the binarized image memory and the character processing means are provided in the optical character reading device, the processing for detecting or recognizing the presence or absence of a character can be performed in a predetermined area as in the first invention. Is limited to only the background pattern.
And the second binarized slice value can be accurately obtained, the accuracy of character presence detection and recognition is improved, and the processing speed is increased.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of an optical character reading device according to a first embodiment of the present invention.

FIG. 2 is a diagram showing an example of a passbook format.

FIG. 3 is a configuration diagram illustrating an outline of a color image reading unit 10 in FIG. 1;

FIG. 4 is a configuration diagram showing a color processing unit 20 in FIG. 1;

FIG. 5 is a configuration diagram showing a histogram circuit 40 in FIG. 1;

6 is a diagram showing the contents of a histogram memory 43 in FIG.

FIG. 7 is a configuration diagram showing a histogram analyzer 50 in FIG. 1;

FIG. 8 is a diagram showing an outline of an attribute determination circuit 60 in FIG. 1;

9 is a configuration diagram of a filter circuit 70 in FIG.

FIG. 10 is a diagram showing a pixel array.

FIG. 11 is a diagram showing a scanning direction of a reading target 5;

FIG. 12 is a diagram showing an achromatic color histogram and a chromatic color histogram.

FIG. 13 is a diagram showing processing contents of a histogram analysis unit 50;

FIG. 14 is a flowchart illustrating an achromatic color Pareto calculation process.

FIG. 15 is an achromatic maximum component pareto diagram expressed as a percentage corresponding to FIG. 12 (a).

FIG. 16 is a flowchart showing a chromatic pareto calculation process.

FIG. 17 is a chromatic color maximum component Pareto diagram expressed as a percentage corresponding to FIG. 12 (b).

FIG. 18 is a flowchart illustrating an achromatic color Pareto change rate calculation process.

FIG. 19 is a diagram illustrating the achromatic color maximum component Pareto change rate expressed as a percentage corresponding to FIG. 15;

FIG. 20 is a flowchart illustrating a chromatic Pareto change rate calculation process.

21 is a diagram illustrating a chromatic color maximum component Pareto change rate expressed as a percentage corresponding to FIG. 17;

FIG. 22 is a flowchart showing a chromatic color Pareto change rate analysis process (1).

FIG. 23 is a flowchart showing a chromatic color Pareto change rate analysis process (2).

FIG. 24 is a flowchart showing a chromatic color Pareto change rate analysis process (3).

FIG. 25 is a diagram showing a determination condition for attribute determination in step STP49 in FIG. 22.

FIG. 26 is a flowchart showing an achromatic color Pareto change rate analysis process (1).

FIG. 27 is a flowchart showing an achromatic color Pareto change rate analysis process (2).

FIG. 28 is a flowchart showing achromatic color Pareto change rate analysis processing (3).

FIG. 29 is a diagram showing the logic of the combinational logic circuit 77;

FIG. 30 is a configuration diagram of an optical character reading device according to a second embodiment of the present invention.

FIG. 31 is a configuration diagram showing a histogram circuit 40A in FIG. 30;

FIG. 32 is a configuration diagram of an optical character reading apparatus according to a third embodiment of the present invention.

FIG. 33 is a configuration diagram showing a histogram circuit 40B in FIG. 32;

FIG. 34 is a configuration diagram of an optical character recognition device according to a modification of the first embodiment.

FIG. 35 is a configuration diagram showing a binarization circuit 100 in FIG. 34;

[Explanation of symbols]

 5 Reading target 10 Color image reading unit 20 Color processing unit 30 Multi-valued image memory 40, 40A, 40B Histogram circuit 41, 48b Sub-scanning direction counter register 47, 48a Main scanning direction counter register 50 Histogram analysis unit 60 Attribute determination circuit 70 Filter Circuit 80 Binary image memory 90 Print line detection circuit 100 Binary circuit

Claims (5)

[Claims] 1. A background pattern is printed in color on a surface and a character is superimposed on the background pattern and an image is optically acquired from an object to be written or printed in achromatic color to detect or recognize the presence or absence of the character. An optical character reading device, a reading unit that optically acquires an image from the reading target, scans the image, and outputs a color component signal representing each pixel of the reading target in pixel units; Region setting means for setting a range corresponding to an arbitrary number of predetermined regions in the image, determining whether the pixel is a chromatic pixel or an achromatic pixel based on the color component signal, and for each of the set ranges By analyzing the respective appearance frequencies, a first binarized slice value for distinguishing the tint block and the character with respect to the pixel determined to be the chromatic pixel for each of the ranges, Analyzing means for obtaining a second binarized slice value for discriminating the tint block and the character with respect to a pixel determined to be an achromatic pixel; and using the first and second binarized slice values. And a post-processing means for detecting or recognizing the presence / absence of the character in each of the predetermined regions of the reading target based on an image obtained as a result of the binarization. Optical character reader. 2. A copy-forgery-inhibited pattern is printed in color on the surface, and a character is superimposed on the copy-forgery-inhibited pattern and an image is optically obtained from an object to be written or printed in achromatic color. In the optical character reading device to perform, the image of the reading target is optically acquired, and the main scanning and the sub-scanning are performed on the image of the reading target to perform red, green, and blue representing each pixel of the reading target. A color image reading means for outputting a primary color component signal on a pixel-by-pixel basis; and a color identification signal of the identification result by identifying whether the pixel is a chromatic pixel or an achromatic pixel based on the red, green and blue primary color component signals. And a color processing means for outputting a signal having the highest luminance among the red, green and blue primary color component signals as a multi-valued luminance signal, and a multi-processing means for storing the color identification signal and the luminance signal. Value An image memory; an area setting means for setting a range corresponding to an arbitrary number of predetermined areas in the reading object in the sub-scanning direction of the image; and inputting the color identification signal and the luminance signal, For each range, for each of the chromatic color pixels and the achromatic color pixels, a histogram acquisition unit that counts the number of pixels entering each gradation level of the luminance signal to obtain a chromatic color histogram and an achromatic color histogram; A first binarized slice value for analyzing the unevenness of the histogram and distinguishing the tint block and the character for the pixel determined to be the chromatic pixel, and the tint block for the pixel determined to be the achromatic pixel. Histogram analysis means for obtaining, for each set range, a second binarized slice value for distinguishing the character from the character. The color identification signal, the luminance signal, the first binarized slice value, and the second 2
A pixel-slicing value is input, the pixel-of-interest is sequentially determined whether the pixel of interest belongs to a pixel of white, black, or another color, and the determination result is output as a pixel attribute. A pixel attribute is inputted, and the pixel of interest forms a pixel forming part of the character or another part based on the pixel attribute of the pixel of interest or the relation between the pixel of interest and the pixel attribute of surrounding pixels. Filter means for making the pixels constituting part of the character black and the pixels constituting the other parts white, and a binarization result obtained by the filter means A binarized image memory for accumulating as a binarized image, and a character processing unit for detecting or recognizing the presence or absence of a character in each of the predetermined regions to be read based on the binarized image stored in the binarized image memory And characterized by having Optical character reader for. 3. A color image reading unit, a color processing unit, and a multi-valued image memory according to claim 2, wherein a range corresponding to an arbitrary number of predetermined regions in the reading target is set in the main scanning direction of the image. 3. An optical character reading apparatus comprising: an area setting unit; and a histogram acquisition unit, a histogram analysis unit, a pixel attribute determination unit, a filter unit, a binarized image memory, and a character processing unit according to claim 2. . 4. A color image reading unit, a color processing unit, and a multi-valued image memory according to claim 2, wherein a range corresponding to an arbitrary number of predetermined regions in the reading target is set in the main scanning direction and sub-scanning of the image. 3. An optical system comprising: an area setting unit for setting a direction; and a histogram acquisition unit, a histogram analysis unit, a pixel attribute determination unit, a filter unit, a binarized image memory, and a character processing unit according to claim 2. Type character reader. 5. A color image reading means, a color processing means, a multi-valued image memory according to claim 2, 3 or 4,
An area setting unit, a histogram acquisition unit, and a histogram analysis unit; inputting the color identification signal and the luminance signal in the set range read from the multi-valued image memory, wherein the color identification signal indicates the chromatic color pixel For a pixel, the luminance signal is binarized to white or black with the first binarized slice value, and for a pixel whose color identification signal indicates the achromatic pixel, the second binarization is performed. 3. A binarizing means for binarizing the luminance signal to white or black with a slice value, and a binarized image memory for storing a binarized result of the binarizing means as a binarized image. An optical character reading device comprising: the character processing means according to claim 3 or 4.