JP2006229817A - Background detection method, program, recording medium, image processing apparatus, and image forming apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a background detection method capable of properly and surely detecting only a background region without the need for using a histogram, and to provide an image processing apparatus or the like employing the method. <P>SOLUTION: The image processing apparatus 20 is provided with a background detection processor 24 or the like. The background detection processor 24 includes a classification means 24A, a discrimination means 24B, and a background detection means 24C. When a difference between an RGB value of a target pixel and an average of RGB values of class information of a class to which adjacent pixels belong is a first threshold or below, the classification means 24A classifies the target pixel to the class to which the adjacent pixels belong, and when the difference exceeds the first threshold, the classification means 24A classifies the target pixel to a newly setting class. The discrimination means 24B discriminates whether or not each class includes pixels located at edges of the whole image and discriminates whether or not each class includes pixels whose number is a first rate or over with respect to the number of all the pixels of the whole image. The background detection means 24C detects classes belonging to the background region on the basis of a result of the discrimination by the discrimination means 24B. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a background detection method for detecting a background area for removing a background from an image, a program for executing the method, a recording medium storing a program for executing the method, an image processing apparatus and an image using the method The present invention relates to a forming apparatus.

  2. Description of the Related Art Conventionally, a technique for creating a histogram for detecting a background area of an image is known.

  For example, based on the input image data, create a histogram showing the density and its frequency, find the maximum density value and minimum density value from the histogram, and calculate the temporary background density from the maximum frequency density and the maximum density value. A technique for detecting the background density from the temporary background density and the minimum density value has been proposed (for example, see Patent Document 1).

  Also, based on the input image data, create a histogram showing the density and its frequency, detect the maximum frequency and the density with the maximum frequency, and remove the background density to be removed from those values A technique for determining a range, comparing the maximum and minimum values of the density range with image data to determine whether the density range is within the range, and outputting image data or background-removed data based on the determination result It has been proposed (see, for example, Patent Document 2).

Further, based on the input color image information, it is determined whether the color image is a character image or a photographic image. If it is determined that the character image is included, the background is removed and the image is a photographic image. A technique has been proposed in which the background removal is not performed when the determination is made (see, for example, Patent Document 3).
JP-A-1-196971 Japanese Patent Laid-Open No. 3-44268 Japanese Patent No. 2984280

  However, if a histogram is simply created and the density at which the frequency is the maximum value is determined as the background density, it may be determined that the background is not actually the background. For example, in an image such as a gradation, up to a certain place in the image may be determined as an image area, and a subsequent area that is actually an image area may be determined as a background.

  Further, in the technique of the above-mentioned Patent Document 1, since the entire image data is uniquely processed with a threshold value based on the detected background density, a portion that is not actually a background may be removed as a background. is there.

  Similarly, in the technique of Patent Document 2, since the entire image data is uniquely processed with a threshold based on the determined background density, for example, pixels having the same density as the determined background density are included in a part of the photographic image. If present, a part of the photographic image may be removed as a background.

  Furthermore, in the technique of Patent Document 3, if it is determined that the image is a photographic image, the background is not removed, but if a pixel having the same density as the background exists in the character image, the background is removed. The problem has not been solved.

  An object of the present invention is to provide a background detection method and an image processing apparatus using the same, which can detect only a background area appropriately and accurately without using a histogram.

  The background detection method of the present invention is configured as follows to solve the above-described problems.

  (1) A difference value between the pixel value of the target pixel in the entire image based on the image data to be processed and the class pixel value based on the pixel values of all the pixels belonging to the class to which the adjacent pixel adjacent to the target pixel belongs. A first step of calculating, and when the difference value calculated in the first step is less than or equal to a preset first threshold, classifying the pixel of interest into a class to which the adjacent pixel belongs, and the difference value is A second step of classifying the pixel of interest into a newly set class when the first threshold is exceeded, and after classifying all the pixels of the entire image in the second step, A third step for determining whether or not to include a pixel located at an end, and after all the pixels of the whole image are classified in the second step, each class is a whole pixel of the whole image. A fourth step for determining whether or not the pixel includes a number of pixels equal to or greater than a preset first ratio, and a class that is a base region based on the determination results in the third step and the fourth step And a fifth step of detecting.

  In this configuration, pixels at successive positions and close in pixel value are classified into the same class. Thereby, all the pixels of the entire image are classified into a plurality of classes having different pixel values or separated positions. Then, it is determined whether or not each class includes a pixel located at the end of the entire image, and whether or not each class includes a number of pixels equal to or greater than a first ratio with respect to the total number of pixels of the entire image. . In the determination result, when the class is located at the end of the entire image, or when the number of pixels belonging to the class is larger than a predetermined value, these classes are detected as the background area.

  Here, the class refers to a set of pixels having pixel values close to each other and adjacent to each other in the vertical direction or the horizontal direction.

  (2) The class pixel value of the class to which the target pixel belongs in the second step is updated according to the pixel value of the target pixel.

  In this configuration, since the class pixel value is updated every time a new pixel is classified, the pixel value of the target pixel and the class pixel value are accurately compared.

  (3) After the second step and before the third step and the fourth step, in two adjacent classes, the maximum value of the pixel values of all the pixels belonging to each class, The method further comprises the step of integrating the two classes into a single class based on a result of comparing at least one of a minimum value and an average value of pixel values of all the pixels with each other. .

  In this configuration, the maximum value, minimum value, average value, etc. of pixel values of pixels belonging to each class are compared for two adjacent classes that have been divided into different classes in the process of classifying pixels into classes. Based on the result, the two classes are merged into a single class.

  (4) In the fifth step, a class determined to include a pixel located at an end of the whole image in the third step is referred to as a first background class, and in the fourth step, all the pixels of the whole image When a class determined to include a number of pixels equal to or greater than a first ratio set in advance for a number is referred to as a second background class, a first group consisting of classes corresponding only to the first background class, Each class is classified into a second group consisting of classes corresponding only to the second base class, and a third group consisting of classes corresponding to both the first base class and the second base class. The class belonging to the group having the largest total number of pixels belonging to each of the classes belonging to each of the first to third groups is detected as a background region. .

  In this configuration, after a class that is a candidate for a base region is selected according to two conditions for detection of the base region determined in the third step and the fourth step (first stage), the selected class is Furthermore, classes that are classified into three groups based on the above two conditions and that belong to the group having the largest number of pixels are detected as background regions (second stage). In this way, the base region is detected through two stages of determination.

  (5) When the total number of pixels belonging to all classes detected as the background area in the fifth step is less than a preset second ratio with respect to the total number of pixels of the image, A sixth step of determining that there is no background area in the image is further provided.

  In this configuration, if the total number of pixels belonging to all classes detected as the background area in the fifth step does not exceed a predetermined ratio with respect to the total number of pixels of the image, the entire image includes the background area. Therefore, it is determined that there is no background area.

  (6) A second threshold value in which a subtraction value obtained by subtracting the minimum value from the maximum value among the pixel values of all the pixels belonging to each class detected as the background area in the fifth step is set in advance. A seventh step for determining that the class is not a background area when the average value of all the pixels belonging to each class detected as a background area exceeds or is less than a preset third threshold value. Is further provided.

  In this configuration, even if a class is detected as a background area, if the difference between the maximum and minimum pixel values of pixels belonging to the class is greater than a predetermined value, the class is not a background area but an image area. It is thought that. In addition, when the average value of the pixel values of all the pixels belonging to the class is less than a predetermined value, the class is unlikely to be a background area.

  (7) For a pixel belonging to a class detected as a background area, an eighth step of performing a background removal process for correcting the pixel value of the pixel so that the background is not printed as an image when there is a background is further provided. It is characterized by.

  In this configuration, the background removal process is performed only on the pixels belonging to the class detected as the background area appropriately and accurately by the above-described background detection method. That is, even if a pixel having the same pixel value as the pixel in the background region exists in the character image and the photographic image, the pixel value of the pixel in the character image and the photographic image is not changed. Only the pixel value of the pixel that is the region is corrected.

  The program of the present invention is configured as follows to solve the above-described problems.

  (8) A difference value between the pixel value of the target pixel in the entire image based on the image data to be processed and the class pixel value based on the pixel values of all the pixels belonging to the class to which the adjacent pixel adjacent to the target pixel belongs A first step of calculating, and when the difference value calculated in the first step is less than or equal to a preset first threshold, classifying the pixel of interest into a class to which the adjacent pixel belongs, and the difference value is A second step of classifying the pixel of interest into a newly set class when the first threshold is exceeded, and after classifying all the pixels of the entire image in the second step, A third step for determining whether or not to include a pixel located at an end, and after all the pixels of the whole image are classified in the second step, each class is a whole pixel of the whole image. A fourth step of determining whether or not the pixel includes a number of pixels greater than or equal to a preset ratio, and a class of detecting a class that is a base region based on the determination results in the third step and the fourth step 5 steps are made to perform by a computer.

  In this configuration, pixels at successive positions and close in pixel value are classified into the same class. Thereby, all the pixels of the entire image are classified into a plurality of classes having different pixel values and the like. Then, it is determined whether or not each class includes a pixel located at the end of the entire image, and whether or not each class includes a number of pixels equal to or greater than a first ratio with respect to the total number of pixels of the entire image. . If the class is located at the edge of the whole image in the judgment result, or if a group of pixels with the same pixel value (class) is larger than a predetermined size, the class is detected as a background area. Is done.

  The recording medium of the present invention is configured as follows to solve the above-described problems.

  (9) A difference value between the pixel value of the target pixel in the entire image based on the image data to be processed and the class pixel value based on the pixel values of all the pixels belonging to the class to which the adjacent pixel adjacent to the target pixel belongs A first step of calculating, and when the difference value calculated in the first step is less than or equal to a preset first threshold, classifying the pixel of interest into a class to which the adjacent pixel belongs, and the difference value is A second step of classifying the pixel of interest into a newly set class when the first threshold is exceeded, and after classifying all the pixels of the entire image in the second step, A third step for determining whether or not to include a pixel located at an end, and after all the pixels of the whole image are classified in the second step, each class is a whole pixel of the whole image. A fourth step of determining whether or not the pixel includes a number of pixels greater than or equal to a preset ratio, and a class of detecting a class that is a base region based on the determination results in the third step and the fourth step The computer stores a program that causes the computer to execute the five steps.

  In this configuration, pixels at successive positions and close in pixel value are classified into the same class. Thereby, all the pixels of the image are classified into a plurality of classes having different pixel values and the like. Then, it is determined whether or not each class includes a pixel located at the end of the entire image, and whether or not each class includes a number of pixels equal to or greater than a first ratio with respect to the total number of pixels of the entire image. . If the class is located at the edge of the whole image in the judgment result, or if a group of pixels with the same pixel value (class) is larger than a predetermined size, the class is detected as a background area. Is done.

  The image processing apparatus of the present invention is configured as follows in order to solve the above-described problems.

  (10) The difference value between the pixel value of the target pixel in the entire image based on the image data to be processed and the class pixel value based on the pixel values of all the pixels belonging to the class to which the adjacent pixel adjacent to the target pixel belongs Calculating and classifying the pixel of interest into a class to which the adjacent pixel belongs when the difference value is equal to or less than a preset first threshold, and the pixel of interest when the difference value exceeds the first threshold And classifying means for classifying the pixels into a newly set class, and determining whether each class includes a pixel located at an end of the whole image after all the pixels of the whole image are classified by the classification means. Whether or not each class includes a number of pixels greater than or equal to a preset ratio with respect to the total number of pixels of the whole image after all the pixels of the whole image are classified by the determination unit of 1 and the classification unit The A second determination means for, characterized in that it comprises a base detection means for detecting a class is a background region based on the determination result of the first judging means and the second judging means.

  In this configuration, pixels at successive positions and close in pixel value are classified into the same class. Thereby, all the pixels of the entire image are classified into a plurality of classes having different pixel values and the like. Then, it is determined whether or not each class includes a pixel located at the end of the entire image, and whether or not each class includes a number of pixels equal to or greater than a first ratio with respect to the total number of pixels of the entire image. . If the class is located at the edge of the whole image in the judgment result, or if a group of pixels with the same pixel value (class) is larger than a predetermined size, the class is detected as a background area. Is done.

  (11) For a pixel belonging to a class detected as a background area by the background detection means, a background removal process for performing a background removal process for correcting the pixel value of the pixel so that the background is not printed as an image when there is a background. A means is further provided.

  In this configuration, the background removal process is performed only on the pixels belonging to the class detected as being the background area appropriately and accurately. That is, even if a pixel having the same pixel value as the pixel in the background region exists in the character image and the photographic image, the pixel value of the pixel in the character image and the photographic image is not changed. Only the pixel value of the pixel that is the region is corrected.

  The image forming apparatus of the present invention is configured as follows in order to solve the above-described problems.

  (12) The image processing apparatus according to claim 11, input means for inputting image data to be subjected to background removal processing in the image processing apparatus, and image processing including the background removal processing by the image processing apparatus. And an output means for outputting the applied image.

  In this configuration, the output means outputs a high-quality image from which only the background is properly and accurately removed by the image processing apparatus.

  According to the present invention, the following effects can be obtained.

  (1) After all the pixels are classified into classes each composed of continuous pixels based on the pixel values, the background in class units based on the position of the class and the size of the class (number of pixels belonging to the class) Since it is determined whether or not it is an area, it is possible to appropriately detect only the base area. Further, a pixel in the image area having a pixel value comparable to that of the pixel in the background area is not erroneously detected as the background area. Furthermore, since it is not necessary to create a histogram, the load on the device used for detecting the background area can be reduced.

  (2) Since the pixel value of the target pixel and the class pixel value can be accurately compared, each pixel can be properly classified into a class to which it originally belongs. As a result, only the base region can be detected appropriately and accurately.

  (3) In the process of classifying a pixel into a class, it is divided into different classes, but by integrating two adjacent classes that can be originally made into a single class, it is easy to detect the subsequent background area. Can be.

  (4) Since the background area is detected through two stages of determination, only the background area can be detected more appropriately and accurately.

  (5) Since the background area is considered to be composed of a predetermined number or more of pixels with respect to the total number of pixels of the image, the total number of pixels belonging to the class determined to be the background area is less than the predetermined number. By determining that there is no background area in the image, it is possible to prevent an erroneous area from being detected as the background area. Therefore, it is possible to detect only the base region more appropriately and accurately.

  (6) By excluding a class that is not considered to be a base region from the base region, it is possible to detect only the base region more appropriately and accurately.

  (7) Since the background removal process is performed only on the pixels belonging to the class detected as the background area appropriately and accurately, only the background is removed with high accuracy without degrading the image quality of the character image and the photographic image. Can do.

  (8) A high-quality image can be output based on image data from which only the background has been removed appropriately and accurately.

  Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing a schematic configuration of a color image forming apparatus according to an embodiment of the present invention.

  A color image forming apparatus (corresponding to the image forming apparatus of the present invention) 10 is a color image input apparatus (corresponding to the input means of the present invention) 11, a color image processing apparatus (corresponding to the image processing apparatus of the present invention). .) 20 and a color image output device (corresponding to the output means of the present invention) 12. The image data of RGB (R: red, G: green, B: blue) analog signals input by the color image input device 11 is subjected to processing described later in the color image processing device 20 to obtain CMYK (C: cyan). , M: magenta, Y: yellow, K: black) are output to the color image output device 12 as digital color signals.

  In this embodiment, the color image input device 11 is a scanner having a CCD (Charge Coupled Device), reads a reflected light image from an image of a document as an RGB analog signal, and outputs it to the color image processing device 20. The color image processing apparatus 20 will be described later. The color image output device 12 is a printer such as an electrophotographic system or an inkjet system that outputs image data of an image of a document onto a recording sheet. The color image output device 12 may be a display device such as a display.

  The color image processing apparatus 20 includes an A / D (analog / digital) conversion unit 21. The A / D conversion unit 21 converts RGB analog signals input from the color image input device 11 into, for example, 8-bit digital signals. The RGB digital signals output from the A / D conversion unit 21 are input to the shading correction unit 22. The shading correction unit 22 performs processing for removing various distortions generated in the illumination system, the imaging system, and the imaging system of the color image input device 11 on the input RGB digital signal. The RGB signal output from the shading correction unit 22 is input to the input tone correction unit 23.

  The input tone correction unit 23 adjusts the color balance of the RGB signal (RGB reflectance signal) from which various distortions have been removed by the shading correction unit 22 and is used in the color image processing apparatus 20 such as a density signal. The signal is converted into an easy-to-handle signal. The RGB signals output from the input tone correction unit 23 are temporarily saved in a memory (not shown) and then input to the background detection processing unit 24 pixel by pixel.

  Based on the input RGB signal, the background detection processing unit 24 detects whether or not there is a background area in the entire image, and if there is a background area, the background area is detected. Here, the image of one original is called an entire image, and the entire image is composed of an image area and a background area. The image area refers to an area where a character image or a photographic image is present. The background area refers to an area other than the image area where the background exists. Details of processing performed by the background detection processing unit 24 will be described later.

  When the background detection processing unit 24 detects a background region, the background detection result is input to the background removal processing unit 25 in synchronization with the RGB signal. In the background removal processing unit 25, background removal processing is performed on the image data (RGB signal) based on the background detection result. The RGB signals output from the background removal processing unit 25 are input to the region separation processing unit 26.

  The area separation processing unit 26 separates each pixel in the image area into one of a character area, a halftone dot area, and a photograph area based on the RGB signal. In the region separation processing unit 26, based on the separation result, a region identification signal indicating to which region each pixel belongs is sent to the spatial filter processing unit 27, the black generation and under color removal unit 30, and the gradation reproduction processing unit 32. And the RGB signal input from the background removal processing unit 25 is output to the spatial filter processing unit 27 in synchronization with the region identification signal.

  In the spatial filter processing unit 27, spatial filter processing using a digital filter is performed on the image data of the RGB signal input from the region separation processing unit 26 based on the region identification signal. As a result, the spatial frequency characteristic is corrected, and blurring and graininess deterioration of the output image in the color image output device 12 are prevented. For example, a region separated into character regions in the region separation processing unit 26 is subjected to sharp enhancement processing, which is one of spatial filter processing, in the spatial filter processing unit 27 in order to improve the reproducibility of black characters or color characters. The high frequency component is emphasized. The region separated into halftone dot regions by the region separation processing unit 26 is subjected to low-pass filter processing for removing the input halftone dot component by the spatial filter processing unit 27. The image data processed by the spatial filter processing unit 27 is input to the scaling processing unit 28.

  In the scaling processing unit 28, enlargement calculation processing or reduction calculation processing is performed on the image data on which the spatial filter processing has been performed so as to obtain a predetermined magnification. The processed RGB signal is input to the color correction unit 29.

  In the color correction unit 29, the input RGB signal image data is converted into a CMY color space, and color correction is performed in accordance with the color image output device 12. The corrected CMY signal image data is input to the black generation and under color removal unit 30.

  The black generation and under color removal unit 30 performs a black generation process for generating a black (K) signal from the CMY three-color signals after color correction, and newly subtracts the K signal acquired by the black generation process from the original CMY signal. The CMY three-color signal is converted into a CMYK four-color signal. An example of black generation processing is shown below. For example, in the process of generating black by skeleton black (general process), the input / output characteristic of the skeleton curve is y = f (x), the input data is C, M, Y, and the output data is C ′. , M ′, Y ′, and K ′, and the UCR (Under Color Removal) rate is α (0 <α <1), the black generation and under color removal processing is expressed by the following equation (1).

  The CMYK signal output from the black generation and under color removal unit 30 is input to the gradation reproduction processing unit 32 via the output gradation correction unit 31.

  Similar to the spatial filter processing unit 27, the gradation reproduction processing unit 32 performs predetermined processing on the image data of the CMYK signal based on the region identification signal input from the region separation processing unit 26. For example, the region separated into the character regions in the region separation processing unit 26 is used to reproduce high-frequency components in the color image output apparatus in the tone reproduction processing unit 32 in order to improve the reproducibility of black characters or color characters. A binarization process or a multi-value process is selected and applied as appropriate. The region separated into halftone dot regions by the region separation processing unit 26 is an output for converting a signal such as a density signal into a halftone dot area ratio that is a characteristic value of the color image output device 12 in the output tone correction unit 32. After the gradation correction processing is performed, the gradation reproduction processing unit 32 performs gradation reproduction processing (halftone generation) in which the entire image is finally separated into pixels and each gradation can be reproduced. Applied. Further, the region separated into the photographic regions in the region separation processing unit is subjected to binarization processing or multi-value processing so as to be suitable for gradation reproducibility in the color image output apparatus.

  The image data subjected to each process in the color image processing apparatus 20 described above is temporarily stored in a storage unit (not shown), read out at a predetermined timing, and input to the color image output apparatus 12. The above processing is controlled by a CPU (Central Processing Unit) (not shown).

  In the color image processing apparatus 20, the background detection processing unit 24 and the background removal processing unit 25 are arranged immediately before the region separation processing unit 26, but the present invention is not limited to this. It may be after the spatial filter processing unit 27 or the like, and may be arranged anywhere after the input tone correction unit 23 and before the color correction unit 29. In addition, the background detection processing unit 24 and the background removal processing unit 25 may not be consecutively disposed, and the region separation processing unit 26, the spatial filter processing unit 27, and the like may be disposed therebetween.

  Next, the configurations of the background detection processing unit 24 and the background removal processing unit 25, which are features of the present invention, will be described.

  FIG. 2 is a block diagram illustrating the configuration of the background detection processing unit 24 and the background removal processing unit 25.

  The background detection processing unit 24 includes a classification unit 24A, a determination unit 24B, and a background detection unit 24C. In this embodiment, the determination unit 24B also serves as the first determination unit and the second determination unit of the present invention.

  Image data composed of RGB signals output from the input tone correction unit 23 is input to the classification unit 24A.

  The classification unit 24A calculates a difference value between the pixel value of the target pixel in the entire image based on the image data to be processed and the class pixel value of the class to which the adjacent pixel adjacent to the target pixel belongs, and based on the difference value Then, a classification process for classifying the target pixel into a class to which the adjacent pixel belongs or a newly set class is performed.

  Here, in this embodiment, the pixel value is the value of the RGB signal of the pixel (here, the target pixel), and the class pixel value is the pixel value (RGB signal value) of all the pixels belonging to the class. Average value.

  Class information of each class is held in a memory (not shown). The class information exists at the top of the pixels belonging to the class 40, as shown in FIG. 3, the maximum value / minimum value / average value (class pixel value) of RGB signals of the pixels belonging to the class. Pixel (top pixel) 41, bottommost pixel (bottom pixel) 42, leftmost pixel (leftmost pixel) 43, and rightmost pixel (rightmost pixel) 44). , Including the number of pixels belonging to the class (number of pixels). Note that the class information may further include related class information. Each class is assigned an index number, and an index number of the class is assigned to a pixel belonging to each class.

  The above-described classification unit 24A reads out the class information of the pixels already classified into any class among the adjacent pixels of the target pixel, and determines whether or not the adjacent pixel belongs to the class to which the adjacent pixel belongs. If the classification unit 24A determines that the target pixel belongs to the class to which the adjacent pixel belongs, the classification unit 24A assigns the index number of the class to which the adjacent pixel belongs to the target pixel and updates the class information. If the classifying unit 24A determines that the target pixel does not belong to the class to which the adjacent pixel belongs, the classifying unit 24A classifies the target pixel into the newly set class and also sets a new index number that is not set in the existing class for the target pixel. And class information of the class is generated.

  The class information updated every time each pixel is classified is temporarily saved in the memory (not shown). At this time, the index number of the pixel classified into the class by the classification unit 24A is also temporarily saved in the memory together with the pixel value of the pixel. The class information including the index number may be latched until the background removal process described later can be started.

  After the classification process is completed for all pixels of the entire image based on the image data to be processed, all class information is input to the determination unit 24B.

  The determination unit 24B performs determination processing for determining whether or not a predetermined condition is satisfied based on the position information and the number of pixels included in the class information. The determination result is input to the background detection means 24C.

  The background detection unit 24C determines whether the class is a class corresponding to the background region based on the determination result input from the determination unit 24B (primary determination). Further, the background detection unit 24C further determines whether or not the class really corresponds to the background area based on the maximum value / minimum value / average value of the class information of the class determined to correspond to the background area in the primary determination. Determine (secondary determination).

  The class information of the class determined to correspond to the background area by the background detection means 24C and the pixel values of all the pixels belonging to the class are output from the background detection processing unit 24 as the background detection result and are supplied to the background removal processing unit 25. Entered.

  The background removal processing unit 25 includes background removal processing means 25A. Based on the class information and the pixel value output as the background detection result, the background removal processing unit 25A performs a predetermined background removal process on the RGB signal of the pixel determined to be the background area. The background removal processing refers to processing for making a pixel white by increasing each value of the RGB signal of the pixel determined to be a background region, for example. The image data that has undergone the background removal processing is input to the region separation processing unit 26.

  FIG. 4 is a flowchart showing a processing procedure in the classification means 24A of the background detection processing unit 24.

  First, the classification unit 24A reads out image data of RGB signals output from the input tone correction unit 23 and saved in the memory for one pixel (S1). Here, the input R, G, and B image data are 8-bit data and take values from 0 to 255.

  Next, the classification unit 24A determines whether or not the position of the read pixel (target pixel) is the upper left pixel of the image (S2). If the classifying unit 24A determines that the pixel of interest is the upper left pixel, it sets a new class and assigns a new class index number (for example, the minimum value “0” as the index number) to the pixel of interest. The position information (coordinate value) of the target pixel is set as the position information of the leftmost pixel of the class to which the target pixel belongs, and the maximum, minimum, and average values of the R, G, and B values of the class are set. Then, R, G, and B values of the pixel of interest are set, and 1 is set as the count value of the class (S3). Then, the classification unit 24A returns to S1 and reads the image data of the next one pixel.

  If the classifying unit 24A determines in S2 that the target pixel is not the upper leftmost pixel, it determines whether or not the target pixel is a pixel on the first line of the entire image (S4). If the classifying unit 24A determines that the pixel of interest is a pixel on the first line, the classifying unit 24A acquires class information to which an adjacent pixel already classified into the class that exists on the left side by one pixel belongs (S5).

  The classifying unit 24A calculates an absolute value of a difference value between each average value of the RGB values of the class information and the RGB value of the target pixel (corresponding to the first step of the present invention), and based on the absolute value. Then, it is determined whether or not the target pixel belongs to an existing class to which the left adjacent pixel belongs (S6). Specifically, when the classification unit 24A determines that the absolute value of each difference value of the RGB values is equal to or less than a first threshold value set in advance, the pixel of interest belongs to the class to which the left adjacent pixel belongs. If the absolute value of the difference value exceeds the first threshold value, it is determined that the target pixel does not belong to the class to which the left adjacent pixel belongs.

  If the classification unit 24A determines that the target pixel belongs to the class to which the left adjacent pixel belongs, the classification unit 24A assigns the index number of the same class as the adjacent pixel to the target pixel, and the position of the right end pixel in the class information of the class The information, maximum value / minimum value / average value, and count value are updated (S7). Then, the classification unit 24A returns to S1 and reads the image data of the next one pixel.

  On the other hand, if the classifying unit 24A determines that the target pixel does not belong to the class to which the left adjacent pixel belongs, a new class is set, and the target pixel is classified into the newly set class (S8). Further, the classification means 24A sets the position information (coordinate value) of the target pixel as the position information of the leftmost pixel of the class to which the target pixel belongs, as in S3, and class R, G, B The R, G, and B values of the pixel of interest are set as the maximum value, minimum value, and average value of the values, respectively, and 1 is set as the class count value. Then, the classification unit 24A returns to S1 and reads the image data of the next one pixel.

  The processes of S7 and S8 correspond to the second step of the present invention.

  If the classifying unit 24A determines in S4 that the target pixel is not the pixel in the first line, it determines whether or not the target pixel is the leftmost pixel in the entire image (S9). If the classifying unit 24A determines that the target pixel is the leftmost pixel of the entire image, the classifying unit 24A acquires class information of a class to which the adjacent pixel above the target pixel belongs (S10). Then, the classification means 24A calculates the difference value of the RGB values by the same process as S6 (corresponding to the first step of the present invention), and the upper adjacent pixel belongs based on the absolute value of the difference value. It is determined whether or not the target pixel belongs to the class (S11).

  If the classification unit 24A determines that the pixel of interest belongs to the class to which the upper adjacent pixel belongs, the classification unit 24A updates the class information of the class to which the pixel of interest belongs, similarly to S7 (S12). Then, the classification unit 24A returns to S1 and reads the image data of the next one pixel. If the classifying unit 24A determines that the target pixel does not belong to the class to which the upper adjacent pixel belongs, similarly to S8, the classifying unit 24A sets a new class and classifies the target pixel into the newly set class ( S13). Then, the classification unit 24A returns to S1 and reads the image data of the next one pixel.

  The processes of S12 and S13 correspond to the second step of the present invention.

  If the classifying unit 24A determines in S9 that the pixel of interest is not the leftmost pixel of the entire image, the class information of the class to which the adjacent pixel above the pixel of interest belongs, and the class information of the class to which the left adjacent pixel belongs Are read from the memory (S14). Then, the classification means 24A calculates the difference value of the RGB values by the same process as S6 (corresponding to the first step of the present invention), and the upper adjacent pixel belongs based on the absolute value of the difference value. It is determined whether the pixel of interest belongs to at least one of the class and the class to which the left adjacent pixel belongs (S15).

  If the classification unit 24A determines in S15 that the pixel of interest does not belong to either the class to which the upper adjacent pixel belongs or the class to which the left adjacent pixel belongs, similarly to S8, the classification unit 24A sets a new class, The target pixel is classified into the class set to (S16), and the process proceeds to S22.

  If the classification unit 24A determines in S15 that the pixel of interest belongs to at least one of the class to which the upper adjacent pixel belongs and the class to which the left adjacent pixel belongs, the upper adjacent pixel further belongs. It is determined whether or not the pixel of interest belongs to both the class and the class to which the left adjacent pixel belongs (S17).

  When the classification unit 24A determines in S17 that the pixel of interest does not belong to both classes, that is, the pixel of interest belongs to only one of the class to which the upper adjacent pixel belongs and the class to which the left adjacent pixel belongs As in S7, the class information of the class to which the target pixel belongs is updated (S18), and the process proceeds to S22.

  The processes of S16 and S18 correspond to the second step of the present invention.

  When the classifying unit 24A determines that the target pixel belongs to both the class to which the upper adjacent pixel belongs and the class to which the left adjacent pixel belongs, the class having the smaller index number of the two classes with respect to the target pixel And the class information of one class to which the pixel of interest belongs is updated (S19).

  Then, the classifying unit 24A calculates a difference value by comparing at least one of the maximum value, the minimum value, and the average value from the class information of the two classes, and calculates the difference value and a predetermined threshold value set in advance. Based on the comparison result, it is determined whether or not the two classes are related classes (S20). The process of S20 corresponds to the step of integrating the two classes of the present invention into a single class.

  If the difference value is equal to or smaller than a predetermined threshold value, the classification unit 24A determines that the above two classes are in a relationship to be a single class (related classes) (S21). Then, the classification unit 24A uses the index number of the class to which the pixel of interest belongs as the related class information in the class information of the class to which the pixel of interest does not belong, of the two classes. Set. By setting the related class information, it becomes possible to determine that the two classes have a relationship to be a single class in later processing. The classification unit 24A proceeds to the process of S22 after the process of S21.

  If the classification unit 24A determines that the two classes are not related classes in S20, the classification unit 24A proceeds to the process of S22.

  The classification unit 24A determines whether or not the classification process has been completed for all the pixels of the entire image, that is, whether or not all the pixels have been classified into classes (S22). If the classification unit 24A determines that all pixels have been classified into classes, the classification unit 24A proceeds to the determination process illustrated in FIG. On the other hand, when the classification unit 24A determines that all the pixels have not been classified into the class, the classification unit 24A returns to S1 and reads the image data of the next one pixel.

  The index number assigned to each pixel and the class information of each class are temporarily saved in the above memory.

  FIG. 5 is a flowchart illustrating a processing procedure in the determination unit 24B of the background detection processing unit 24. First, the determination unit 24B reads out the class information created by the classification process from the memory in order from the largest index number (S31).

  When the index number is set in the related class information in the read class information, the determination unit 24B determines the class information of the class of the index number set in the related class information and the class currently being focused on. Based on the class information, the class information of the class of the index number set in the related class information is updated so as to integrate the two classes into a single class (related class integration process) (S32). The updated class information is saved in the above memory.

  At this time, the determination unit 24B updates each position information so that the area indicated by the position information is expanded. The determination unit 24B compares the maximum values and the minimum values of the two classes with each other and updates them to appropriate values, and updates the average values of the RGB values with two values. Based on the average value and count value of each class, the average value when the two classes are integrated is recalculated and updated. The determination unit 24B also updates the count value to a value obtained by adding the count values of both classes.

  After that, the determination unit 24B sets the position information of the class information of the currently focused class to the same value as the position information of the related class, and sets the maximum value / minimum value / average value of the RGB values to the maximum of the RGB values of the related class. Set to the same value as the value, minimum value, and average value. Further, the determination unit 24B updates the count value of the currently focused class to “0”.

  Next, the determination unit 24B determines whether or not the processing of S31 and S32 is completed for all classes, and repeats the processing of S31 and S32 until the processing is completed for all classes (S33).

  Next, the determination unit 24B reads the updated class information from the memory in order from the smallest index number (S34). Then, the determination unit 24B determines whether or not the class is positioned at the end of the entire image by determining whether or not the pixel includes the pixel positioned at the end of the entire image based on the updated position information of the class information. (S35) (corresponds to the third step of the present invention). If the determination unit 24B determines that the class is located at the end of the entire image, the determination unit 24B adds a flag indicating that it is designated as the first background class to the class information (S36). If the determination unit 24B determines that the class is not located at the end of the entire image, the determination unit 24B proceeds to the process of S37 without adding the flag of the first background class.

  Next, the determination unit 24B uses the count value of the class information to determine whether or not the class of interest includes a number of pixels that is equal to or greater than a preset first ratio with respect to the total number of pixels of the entire image. Determination based on (S37) (corresponds to the fourth step of the present invention). If the determination unit 24B determines that the class of interest includes a number of pixels that is equal to or greater than the first ratio set in advance with respect to the total number of pixels of the entire image, the determination unit 24B is designated as the second background class in the class information. A flag indicating that this is assigned (S38). If the determination unit 24B determines that the class of interest does not include a number of pixels equal to or greater than a preset first ratio with respect to the total number of pixels in the entire image, the determination unit 24B adds a flag for the second background class. Then, the process proceeds to S39.

  The determination unit 24B determines whether or not the processing from S34 to S37 has been completed for all classes, and repeats the processing from S34 to S37 until it has been completed for all classes (S39).

  If the determination unit 24B determines that the processing from S34 to S37 has been completed for all classes, the determination unit 24B ends the determination processing and proceeds to background detection processing.

  FIG. 6 is a flowchart showing a processing procedure in the background detection means 24C of the background detection processing unit 24.

  The background detection means 24C includes a first group consisting of classes corresponding only to the first background class, a second group consisting of classes corresponding only to the second background class, and the first background class and the second group. Each class is classified into a third group consisting of classes corresponding to both of the base classes (S41). Then, the background detection unit 24C detects the class belonging to the group having the largest total count value of the classes belonging to each of the first to third groups as a background region (S42).

  The processing of S41 and S42 corresponds to the fifth step of the present invention.

  The background detection means 24C calculates the sum of the count values of all classes detected as being the background area (S43). Then, the background detection unit 24C determines whether or not the total of the count values is equal to or greater than a second ratio set in advance with respect to the total number of pixels of the entire image (S44).

  When it is determined that the total count value is less than the second ratio with respect to the total number of pixels of the entire image, the background detection unit 24C determines that there is no background area in the entire image (S45). The detection process ends.

  The processes of S44 and S45 correspond to the sixth step of the present invention.

  If the background detection unit 24C determines that the sum of the count values is equal to or greater than the second preset ratio with respect to the total number of pixels of the entire image, the maximum RGB value of each class detected as the background region The minimum value and the average value are read (S46). Then, the background detection means 24C re-determines whether each class detected as the background area is really a background area (S47). Specifically, the background detection unit 24C has a subtraction value obtained by subtracting the minimum value from the maximum value of the RGB values of each class equal to or less than a preset second threshold value, and the average value is the third value. It is determined whether or not the threshold value is exceeded.

  If the subtraction value is equal to or smaller than the second threshold value and the average value is equal to or greater than the third threshold value, the background detection unit 24C determines that the class is finally a background region (S48). If the subtraction value exceeds the second threshold value or the average value is less than the third threshold value, the background detection unit 24C determines that the class is not a background region (S49). The result of the background detection process is saved in the memory.

  The processing from S47 to S49 corresponds to the seventh step of the present invention.

  FIG. 7 is a flowchart showing a processing procedure in the background removal processing unit 25. The background removal processing unit 25 reads from the memory the class information of the class finally determined to be the background area in S47 of the background detection process (S51). Next, the background removal processing unit 25 reads the RGB value of one pixel together with the index number indicating the class assigned in the classification process (S52).

  The background removal processing unit 25 determines whether or not the index number assigned to the pixel of interest is in the index number of the class finally determined to be the background area (S53).

  If the index number assigned to the pixel of interest is within the index number of the class that is finally determined to be the background region, the background removal processing unit 25 performs predetermined background removal processing on the pixel of interest. (S54). Specifically, for example, the background removal processing unit 25 removes the background by setting all the RGB values of the target pixel to 255.

  The process of S54 corresponds to the eighth step of the present invention.

  If the index number assigned to the pixel of interest is not in the index number of the class that is finally determined to be the background region, the background removal processing unit 25 performs background removal processing on the pixel of interest. Instead, the process proceeds to S55.

  The background removal processing unit 25 determines whether or not the processing of S52 to S54 has been completed for all the pixels of the entire image (S55). If it is determined that the processing has been completed, the background removal processing ends, and the result is obtained as a result. The image data composed of the R, G, and B values is output to the region separation processing unit 26. If the background removal processing unit 25 determines that the processing has not been completed for all the pixels, the background removal processing unit 25 repeats the processing of S52 to S54.

  Next, the case where the above-described classification process, determination process, background detection process, and background removal process are performed on a part of the image area 50 in the entire image shown in FIG. 8 will be described in detail. To do. FIG. 8 is an explanatory diagram illustrating an example of an entire image based on image data to be processed. FIG. 9 is an explanatory diagram showing the image data of each pixel in the partial image region 50 shown in FIG. 8, and FIG. 9A shows the image data of each pixel in a state where the classification process is finished up to the fourth pixel. FIG. 9B is an explanatory diagram illustrating image data of each pixel in a state where the classification process has been completed for all the pixels in the image region 50, and FIG. 9C illustrates one pixel. It is explanatory drawing which shows the example of description of the image data of a minute.

  Here, the processing for the pixels starts from the first line from the top, and proceeds downward in order, such as the first line, the second line, the second line, and the third line. . In the same row, the processing starts from the leftmost pixel and proceeds in order toward the right side.

  FIG. 9 is a diagram showing a part of FIG. 8, and therefore there is normally class information of pixels to which index numbers have already been assigned in the vicinity, but here, in order to simplify the description, Assuming that the class information of the class to which the peripheral pixels of the region 50 belong is ignored and only the pixel of FIG. 9 exists, the classification process for the three pixels 51 to 53 will be described as an example. Note that the upper left pixel position will be described as {x, y} = {0, 0}.

  The pixel 51 is an example for explaining the classification process at the stage where only the left adjacent pixel has already been subjected to the classification process and is given a class index number. The maximum value, minimum value, and average value of the RGB values of the class of index number 0 at this stage are as shown in FIG. The adjacent pixel on the left side of the pixel 51 belongs to the class with index number 0. In the classification process of the pixel 51, the RGB value of the pixel 51 is compared with the average value of the RGB values of the class of index number 0. Specifically, as shown in [Equation 2], the absolute value of each difference value of the RGB values is calculated, and the calculation result is compared with the first threshold value.

  Here, for example, “5” is set in advance as the first threshold value. Then, in this case, since the absolute value of all the difference values of the R, G, and B values is equal to or smaller than the threshold value, the pixel 51 is determined to belong to the class to which the left adjacent pixel belongs. Therefore, the index number Idx = 0 of the pixel 51 is set, and the maximum value / minimum value / average value of the RGB value of the class of the index number 0 is updated in consideration of the RGB value of the pixel 51.

  In the example of the pixel 51, there is no change in the maximum value, and R = 247, G = 244, and B = 242. The minimum value is updated to R = 240, G = 240, and B = 238. The average value is updated to R = (245 × 4 + 240) ÷ 5 = 244, G = (242 × 4 + 240) ÷ 5 = 242, and B = (240 × 4 + 238) ÷ 5 = 240. Also, the count value is incremented by 1 and becomes 5. Only the right of the upper, lower, left and right position information is updated, and {Up, Down, Left, Right} = {0, 0, 0, 4}.

  Note that the average value of the RGB values of the class to which the adjacent pixel belongs is not limited to the RGB value of the pixel 51 of interest, and the maximum value or the minimum value can also be compared.

  Next, the classification process of the pixels 52 will be described. The pixel 52 is an example for explaining the classification process at the stage where only the upper adjacent pixel has already been subjected to the classification process and is assigned a class index number. The maximum value, the minimum value, and the average value of the RGB values of the class to which the adjacent pixel above the pixel 52 belongs, that is, the class having the index number 0, are updated in the process for the pixel 51. In the classification process of the pixel 52, the RGB value of the pixel 52 is compared with the average value of the updated RGB values of the class of index number 0 in the same manner as in the classification process for the pixel 51. The absolute values of the difference values of the RGB values are as shown in [Equation 3].

  In the classification process of the pixel 52, the absolute value of the difference value of the R value is larger than the first threshold (here, 5). For this reason, it is determined that the pixel 52 belongs to a class different from the class to which the upper adjacent pixel belongs. Therefore, 1 is added to the maximum value (here, 0) of the index number Idx so far, and Idx = 1 is set, and the maximum value / minimum value / average value of the RGB values of index number 1 are set. Here, since there is only one pixel belonging to the class of index number 1, maximum value = minimum value = average value, and R = 250, G = 239, and B = 245 are set as the respective RGB values. Further, 1 is set as the count value, and {upper, lower, left, right} = {1, 1, 0, 0} is set as the position information.

  Next, the classification process of the pixels 53 will be described. The pixel 53 is an example for explaining the classification process at the stage where the upper adjacent pixel and the left adjacent pixel have already been subjected to the classification process and have been assigned a class index number. The maximum value, the minimum value, and the average value of the RGB values of the class to which the adjacent pixel on the upper side of the pixel 53 belongs, that is, the class with the index number 0, are updated by the processing for the pixel 51, and the adjacent pixel on the left side of the pixel 53 The maximum value, the minimum value, and the average value of the RGB values of the class to which 52 belongs, that is, the class of index number 1, are in a state set by the processing of the pixel 52.

In the classification process of the pixel 53, the RGB value of the pixel 53, the average value of the updated RGB value of the class of index number 0, and the average value of the RGB value of the class of index number 1 are updated. The comparison is performed in the same manner as in the classification processing for 51. The absolute values of the difference values of the RGB values are as shown in [Equation 4] and [Equation 5].

  In [Equation 4], the absolute value of each difference value of RGB values is equal to or less than the first threshold value, whereas in [Equation 5], the absolute value of the R value difference value exceeds the first threshold value. Therefore, it is determined that the pixel 53 belongs to the class to which the upper adjacent pixel belongs. Therefore, the index number Idx = 0 is set for the pixel 53, and the maximum value, the minimum value, and the average value of the RGB values of the index number 0 are updated. The maximum value is unchanged, R = 247, G = 244, B = 242, and the minimum value is updated to R = 240, G = 239, and B = 238. The average value is updated to R = (244 × 5 + 243) ÷ 6 = 244, G = (242 × 5 + 239) ÷ 6 = 242, and B = (240 × 5 + 242) ÷ 6 = 240. Also, the count value is incremented by 1 to 6. The position information is updated only in the top, bottom, left, and right, and {Up, Down, Left, Right} = {0, 1, 0, 4}.

  By performing the classification process as described above for all the pixels, a result as shown in FIG. 9B is obtained. Then, when the classification process is completed for all the pixels, the class information of each class is as shown in FIG.

  Next, the determination process will be described.

  FIG. 11A is an explanatory diagram showing class information of each class at the stage where the classification process has been completed for all the pixels in the partial image region 50 shown in FIG. 8, and FIG. It is explanatory drawing which shows the class information of each class in the stage which the determination process containing a class integration process was complete | finished.

  First, class information of each class is read in descending order of index number. Here, since index number 3 is the largest, it is read from the class information of the class of index number 3.

  Next, it is confirmed whether or not a related class is set in the read class information. Here, the related class = 2 is set in the class information of the class of index number 3. When the related class is set, the class information of the class with the index number 2 set is updated using the class information of the class with the index number 3.

  First, position information is updated. Compare the position information of the class with index number 2 and the position information of the class with index number 3, and select the smaller value for the upper and left, and the larger value for the lower and right so that the area is expanded. Updated. The updated position information is shown as class information of the class of index number 2 in FIG. Here, “left” of the position information is updated from 3 to 2.

  Next, the minimum value / maximum value / average value of the RGB values are updated. For the maximum value, the maximum values of the respective class information are compared, and the larger RGB value is selected. For the minimum value, the minimum value of each class information is compared, and the smaller RGB value is selected. The average value is updated by dividing the value obtained by multiplying the average value of each class information by the count value of each class and dividing by the sum of the count values of the two classes. The average value can be calculated. The maximum value, minimum value, and average value of the updated RGB values are shown as class information of the class of index number 2 in FIG.

  The count value is updated by adding the count values of the two classes. The count value 3 of the index number 3 is added to the count value 5 of the index number 2, and the updated count value is 8.

  Then, the maximum value, minimum value, average value, and position information of the RGB value of the class information of the class of index number 3 are updated to the same values as the class information of the class of index number 2, and the count value of the class information of index number 3 Is updated to 0.

  Next, attention is focused on the class information of the class with index number 2. Since the related class is not set for the class of index number 2, the class information of the class of index number 2 is not updated here.

  Next, attention is focused on the class information of the class with index number 1. Since the related class = 0 is set in the class information of the class with index number 1, update processing is performed in the same manner as in the case of class with index number 3. The result of the update process is shown as class information of the classes of index number 0 and index number 1 in FIG.

  Finally, attention is paid to the class information of the class with index number 0. Since the related class is not set in the class information of the class with index number 0, the class information of the class with index number 0 is not updated here.

  Next, in order from the smallest index number, only the updated class information of each class whose count value is not 0 is read. Then, based on the position information of the class information, it is determined whether or not the class is located at the end of the entire image.

  The image data described here is the image data of a part of the entire image area 50 shown in FIG. 8, and the image area 50 is not actually located at the end of the entire image. The description will be made on the assumption that the class with index number 0 is continuous from the left end of the entire image shown in FIG. 8 to the image region 50.

  In this case, since index number 0 is a class located at the end of the entire image, a flag indicating the first background class is added to the class information of the class of index number 0.

  Further, the count value of the class with index number 0 is compared with a number corresponding to a first ratio set in advance with respect to the number of pixels of the entire image. Actually, since the entire entire image is a target of processing, a large value should be set. However, for convenience of explanation, the entire image is considered in consideration of the size of the image area 50 of interest. The number corresponding to the first ratio with respect to the number of pixels is set to 10 and will be described. Then, the count value of the class with index number 0 is 12, which is larger than the number corresponding to the first ratio, so a flag indicating the second background class is added.

  Next, the class information of the class with index number 2 is read and the position information is confirmed. Then, since the class with index number 2 is not located at the end, it is determined that it is not the first base class. Further, since the count value is a value less than 10 which is a number corresponding to the first ratio, it is determined that the count value is not the second background class. In addition, since there is no class whose count value is not 0, the determination process is terminated and the background detection process is performed.

  Next, background detection processing will be described. In the background detection process, first, a first group consisting of classes corresponding only to the first background class, a second group consisting of classes corresponding only to the second background class, the first background class and the second group Each class is classified into a third group consisting of classes corresponding to both the base class. Next, the count values of the classes classified into the first to third groups are summed for each group, and the sum of the count values is compared between the groups. Then, the class belonging to the group having the largest total count value is determined to be the base area.

  In the above example, it is determined that only the class with index number 0 corresponds to both the first base class and the second base class, so the total count value of the third group is 12. Since there is no class corresponding to only the first background class and no background class corresponding only to the second background class, the sum of the count values of the first group and the second group is zero. Accordingly, since the sum of the count values of the third group is the largest, it is determined that the class belonging to the third group, that is, the class having the index number 0 is the background area.

  Next, the sum of the count values of a single class or a plurality of classes determined to be the background area is calculated. In the example of FIG. 11B, since the class determined to be the background area is only the class with index number 0, the total count value is 12.

  Next, it is determined whether or not the total count value is equal to or greater than a second ratio set in advance with respect to the total number of pixels of the entire image. Here, for convenience of explanation, 10 is set as a number corresponding to the second ratio with respect to the total number of pixels of the entire image, corresponding to the total number of pixels in the region 50. In reality, since the number of pixels belonging to the class of index number 0 should be large in the entire area of the entire image, a number larger than 10 is set.

  In this case, since the total count value “12” is equal to or greater than the number “10” corresponding to the second ratio with respect to the total number of pixels of the entire image, it is determined that the background area exists. This process prevents an erroneous area from being detected as a base area.

  Next, the maximum value, the minimum value, and the average value of the RGB values of the class determined to be the background area (here, the class with index number 0) are acquired. Then, the subtraction value obtained by subtracting the minimum value from the maximum value of each of the RGB values is compared with a preset second threshold value. Here, it is assumed that R, G, and B are set to “30” as the second threshold value.

  At this time, (maximum value of R) − (minimum value of R) = 250−240 = 10, (maximum value of G) − (minimum value of G) = 244−235 = 9, (maximum value of B) − (Minimum value of B) = 250−238 = 12, and both are smaller than the threshold value 30.

  Further, the average value of each of the RGB values of the class determined as the background area is compared with a preset third threshold value. Here, it is assumed that “220” is set as the third threshold value for both RGB. In the example shown in FIG. 11B, the average values of the RGB values of the class with index number 0 are all greater than or equal to the third threshold value 220.

  Since the subtraction value is less than or equal to the second threshold value and the average value is greater than or equal to the third threshold value, the class with index number 0 is finally detected as being a background region, and the background detection processing is terminated. The index number and the like of the class finally determined to be the background area are saved in the memory.

  Note that the threshold value and the ratio used in the background detection processing of this embodiment may be individually set for each of RGB in accordance with the characteristics of the scanner, the density difference and density value of the range where the background is desired to be removed.

  Next, the background removal process will be described. FIG. 12A is an explanatory diagram illustrating an example of image data before the background removal process, and FIG. 12B is an explanatory diagram illustrating an example of image data after the background removal process.

  In the background removal process, RGB values and index numbers are read out pixel by pixel. It is determined whether the image data for one pixel is a base area based on the index number. In the above-described example, the class having the index number “0” is finally detected as the base area. In addition, since the class of index number 1 is set to 0 as the related class, it is substantially integrated into the class of index number 0, and the class of index number 1 is processed in the same manner as the class of index number 0. To do. That is, when the index number assigned to each pixel is 0 or 1, the pixel is determined to be a pixel in the base area.

  When it is determined that the pixel of interest is the background area, each of the RGB values is replaced with 255. As a result, the base is removed. In this example, the value is replaced with 255. However, the value is not limited to this value, and individual optimum values may be set for R, G, and B in consideration of engine performance and the like.

  If it is determined that the pixel of interest is not a background area, the RGB values of the image data are output as they are. Then, the above-described processing is repeated for all the pixels, and image data as shown in FIG. 11B is obtained. When the background removal process is completed for all pixels, the background removal process is completed. The resulting RGB signal image data is output to the region separation processing unit 26.

  According to the image processing apparatus 20, pixels whose density values (RGB values) are close to each other are classified as one class, and after all the pixels are classified into classes, the position of the class and the size of the class with respect to the entire image are classified. Based on the size (number of pixels belonging to the class), it is determined whether or not it is a background area in class units. Therefore, a pixel in an image area having a density value similar to that of a pixel in the background area is erroneously set as the background area. There is no detection. Therefore, it is possible to detect only the ground region appropriately and accurately. In addition, since it is not necessary to create a histogram, the load on the device used for detecting the base region can be reduced.

  In addition, since the class information of the class is updated every time the pixel of interest is classified into a class, the RGB value of the pixel of interest can be accurately compared with the average value of the RGB values of the class information. For this reason, each pixel can be appropriately classified into a class to which it originally belongs.

  In addition, when a related class is set in the class information, the two adjacent classes of the class set as the related class and the class specified as the related class are integrated to perform subsequent background detection processing. Can be easily.

  In addition, after grouping based on the position of the class in the entire image and the number of pixels belonging to the class, the group with the largest number of pixels is detected as the background region, so that the background region is detected through two stages of determination. Thus, only the base region can be detected more appropriately and accurately.

  By the way, it is considered that the base region is composed of a number of pixels equal to or larger than a predetermined ratio with respect to the total number of pixels of the entire image. Therefore, if the total number of pixels belonging to the class determined to be the background area is less than a predetermined ratio with respect to the total number of pixels of the entire image, it is determined that the image has no background area. It is possible to prevent an erroneous area from being detected as an area. Therefore, it is possible to detect only the base region more appropriately and accurately.

  Further, by excluding a class that is not a background area from the background area based on the maximum value, the minimum value, and the average value of the RGB values of the class information, it is possible to detect only the background area more appropriately and accurately. .

  Furthermore, since the background removal processing is performed only on pixels belonging to the class detected as the background area accurately and accurately, it is possible to remove only the background with high accuracy without degrading the image quality of character images and photographic images. it can.

  According to the image forming apparatus 10, it is possible to output a high-quality image based on image data from which only the background has been removed appropriately and accurately.

  The classification processing, determination processing, background detection processing, and background removal processing of the above-described embodiment can be configured as a computer-executable program, and the program can be stored in a recording medium (not shown) so as to be readable by the computer. .

  The recording medium may be a ROM, RAM, a built-in hard disk, or the like that is fixedly provided in the computer main body, or may be a recording medium configured to be separable from the computer main body.

  Recording media configured to be separable from the computer main body include, for example, tape systems such as magnetic tapes and cassette tapes, magnetic disks such as floppy (registered trademark) disks and external hard disks, CD-ROMs, MOs, MDs, DVDs, etc. There are disk systems including optical disks, card systems such as IC cards (including memory cards) and optical cards. In addition, the recording medium configured to be separable from the computer main body is a fixed program including a semiconductor memory such as a mask ROM, an EPROM (Erasable Programmable Read Only Memory), an EEPROM (Electrically Erasable Programmable Read Only Memory), and a flash ROM. It may be a supported medium. Accordingly, a recording medium on which a program for causing a computer to execute the background detection method of the present invention including the above-described classification process, determination process, background detection process, and background removal process can be provided in a portable manner.

  The stored program may be configured to be accessed and executed by a microprocessor, or the program is read and installed in a program storage area (not shown) in the computer. May be configured to execute.

  Furthermore, the image processing apparatus 20 can be configured as a system that can be connected to a communication network including the Internet (not shown), and the recording medium flows so that the image processing apparatus 20 downloads a program via the communication network. Alternatively, it may be a medium carrying a program. When downloading a program via a communication network, the program for executing the download may be stored in the computer main body in advance, or may be installed from another recording medium.

  The program for causing a computer to execute the background detection method described above does not cause only the classification processing, determination processing, background detection processing, and background removal processing of the above-described embodiment to be executed, but is executed in the image processing apparatus 20 illustrated in FIG. It may be configured to be executed comprehensively including other processes.

  As the color image input device 11, for example, a flat bed scanner, a film scanner, a digital camera, or the like is used. As the color image output device 12, for example, an image display device such as a CRT display or a liquid crystal display, or an electrophotographic or inkjet printer that outputs processing results to paper or the like is used. Further, the image forming apparatus 10 may include a modem as a communication unit for connecting to a server or the like via a network.

  Furthermore, in the above-described embodiment, classification processing, determination processing, and background detection processing are performed using RGB values. However, the present invention is not limited to this, and another color space (for example, CMYK or CIELAB) is used depending on detection accuracy, image processing configuration, and the like. Etc.). In this case, in a color space in which density and color information are mixed (such as CMYK), the processing may be performed in the same manner as in RGB values. In a color space in which luminance and chromaticity are separated (such as CIELAB), the luminance is centered. In this process, the chromaticity information may be used as the support information.

1 is a block diagram showing a schematic configuration of a color image forming apparatus according to an embodiment of the present invention. It is a block diagram which shows the structure of a background detection process part and a background removal process part. It is explanatory drawing which shows an example of the collection of the pixels which comprise one class. It is a flowchart which shows the process sequence in the classification | category means of a background detection process part. It is a flowchart which shows the process sequence in the determination means of a background detection process part. It is a flowchart which shows the process sequence in the background detection means of a background detection process part. It is a flowchart which shows the process sequence in a background removal process part. It is explanatory drawing which shows an example of the whole image based on the image data which should be processed. It is explanatory drawing which shows the image data of each pixel in the one part image area | region 50 shown in FIG. 8, (A) is explanatory drawing which shows the image data of each pixel in the state which the classification process was completed to the 4th pixel. (B) is explanatory drawing which shows the image data of each pixel in the state which the classification process was complete | finished about all the pixels of the image area | region 50, (C) is description which shows the description example of the image data for one pixel FIG. It is explanatory drawing which shows the class information of each class in the stage which the classification process was complete | finished about all the pixels of the one part image area shown in FIG. 8, (A) shows the class information of the class of index number 0, (B) shows the class information of the class of index number 1, (C) shows the class information of the class of index number 2, and (D) shows the class information of the class of index number 3. (A) is explanatory drawing which shows the class information of each class in the stage which the classification process was complete | finished about all the pixels of the one part image area | region shown in FIG. 8, (B) includes a related class integration process. It is explanatory drawing which shows the class information of each class in the stage where the determination process was complete | finished. (A) is explanatory drawing which shows an example of the image data before a background removal process, (B) is explanatory drawing which shows an example of the image data after a background removal process.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Image forming apparatus 11 Color image input device (input means)
12 Color image output device (output means)
20 Image processing device 24 Background detection processing unit 24A Classification unit 24B Determination unit (first determination unit, second determination unit)
24C Background detection means 25 Background removal processing section 25A Background removal processing means 40 class

Claims (12)

Calculating a difference value between the pixel value of the target pixel in the entire image based on the image data to be processed and the class pixel value based on the pixel values of all the pixels belonging to the class to which the adjacent pixel adjacent to the target pixel belongs One step,
When the difference value calculated in the first step is less than or equal to a preset first threshold value, the pixel of interest is classified into a class to which the adjacent pixel belongs, and the difference value exceeds the first threshold value A second step of classifying the pixel of interest into a newly set class,
A third step of determining whether or not each class includes a pixel located at an end of the whole image after all the pixels of the whole image are classified in the second step;
After all the pixels of the whole image are classified in the second step, it is determined whether each class includes a number of pixels equal to or more than a first ratio set in advance with respect to the total number of pixels of the whole image. And a fourth step
And a fifth step of detecting a class which is a background region based on the determination results in the third step and the fourth step.   The background detection method according to claim 1, wherein the class pixel value of the class to which the target pixel belongs in the second step is updated according to the pixel value of the target pixel. After the second step and before the third step and the fourth step,
Based on a result of comparing at least one of the maximum value, the minimum value, and the average value of the pixel values of all the pixels in two adjacent classes to each other. The background detection method according to claim 1, further comprising a step of integrating the two classes into a single class. In the fifth step, the class determined to include pixels located at the end of the whole image in the third step is set as a first background class, and the total number of pixels of the whole image in the fourth step. A class determined to include a number of pixels equal to or higher than a preset first ratio is set as a second background class.
A first group consisting of classes corresponding only to the first base class, a second group consisting of classes corresponding only to the second base class, and both the first base class and the second base class A class belonging to a group having the largest total of the number of pixels belonging to each of the classes belonging to each group among the first to third groups. The background detection method according to claim 1, wherein the background detection area is detected as a background area.   If the total number of pixels belonging to all classes detected as a background area in the fifth step is less than a preset second ratio with respect to the total number of pixels of the image, The background detection method according to claim 1, further comprising a sixth step of determining that there is no area.   The subtraction value obtained by subtracting the minimum value from the maximum value of the pixel values of all the pixels belonging to each class detected as the background area in the fifth step exceeds a preset second threshold, or A seventh step of determining that the class is not a background area when the average value of the pixel values of all the pixels belonging to each class detected as the background area is less than a preset third threshold value; The ground surface detection method according to claim 1, further comprising:   For a pixel belonging to a class detected as a background area, an eighth step of performing a background removal process for correcting a pixel value of the pixel so that the background is not printed as an image when there is a background is further provided. The background detection method according to any one of claims 1 to 6. Calculating a difference value between the pixel value of the target pixel in the entire image based on the image data to be processed and the class pixel value based on the pixel values of all the pixels belonging to the class to which the adjacent pixel adjacent to the target pixel belongs One step,
When the difference value calculated in the first step is less than or equal to a preset first threshold value, the pixel of interest is classified into a class to which the adjacent pixel belongs, and the difference value exceeds the first threshold value A second step of classifying the pixel of interest into a newly set class,
A third step of determining whether or not each class includes a pixel located at an end of the whole image after all the pixels of the whole image are classified in the second step;
After all the pixels of the whole image are classified in the second step, it is determined whether each class includes a number of pixels equal to or larger than a preset ratio with respect to the total number of pixels of the whole image. Steps,
A program that causes a computer to execute a fifth step of detecting a class that is a base region based on the determination results in the third step and the fourth step. Calculating a difference value between the pixel value of the target pixel in the entire image based on the image data to be processed and the class pixel value based on the pixel values of all the pixels belonging to the class to which the adjacent pixel adjacent to the target pixel belongs One step,
When the difference value calculated in the first step is less than or equal to a preset first threshold value, the pixel of interest is classified into a class to which the adjacent pixel belongs, and the difference value exceeds the first threshold value A second step of classifying the pixel of interest into a newly set class,
A third step of determining whether or not each class includes a pixel located at an end of the whole image after all the pixels of the whole image are classified in the second step;
After all the pixels of the whole image are classified in the second step, it is determined whether each class includes a number of pixels equal to or larger than a preset ratio with respect to the total number of pixels of the whole image. Steps,
A recording medium storing a computer-readable program for causing a computer to execute a fifth step of detecting a class as a base area based on the determination results in the third step and the fourth step. Calculating a difference value between the pixel value of the target pixel in the entire image based on the image data to be processed and the class pixel value based on the pixel values of all the pixels belonging to the class to which the adjacent pixel adjacent to the target pixel belongs; When the difference value is less than or equal to a preset first threshold, the pixel of interest is classified into a class to which the adjacent pixel belongs, and when the difference value exceeds the first threshold, the pixel of interest is newly Classifying means for classifying into the set class,
First determination means for determining whether or not each class includes a pixel located at an end of the whole image after all the pixels of the whole image are classified by the classification means;
After classifying all the pixels of the whole image by the classification means, a second determination is made as to whether or not each class includes a number of pixels equal to or greater than a preset ratio with respect to the total number of pixels of the whole image. A determination means;
An image processing apparatus comprising: a background detection unit that detects a class that is a background region based on a determination result in the first determination unit and the second determination unit.   For a pixel belonging to a class detected as a background area by the background detection means, when there is a background, a background removal processing means for performing a background removal process for correcting the pixel value of the pixel so that the background is not printed as an image, The image processing apparatus according to claim 10, further comprising: The image processing apparatus according to claim 11,
Input means for inputting image data to be subjected to background removal processing in the image processing apparatus;
An image forming apparatus comprising: an output unit configured to output an image subjected to image processing including the background removal processing by the image processing device.

Images