JP2005109807A - Image processing method, image processor, image forming device and image processing system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processing method for precisely detecting a shading part, and to provide an image processor, an image forming device and an image processing system. <P>SOLUTION: In image data D1 (refer to diagram (a)) obtained by reading a double spread original, a histogram H1 (refer to diagram (b)) is obtained when concentration integrated values obtained by integrating a concentration value on a pixel data string for one line along a Y-axis direction are obtained for all X-coordinate values. Peaks H1a, H1b and H1c in the histogram H1 respectively correspond to a saddle stitch shading part D1a, a right end shading part D1b and a left end shading part D1c. When Fourier transformation is performed on the histogram H1, a spatial frequency distribution F1 (refer to diagram (c)) is obtained. Peaks F1a and F1b are extracted as strong peaks at a low spatial frequency. When inverse Fourier transformation is performed on the peaks, a histogram H2 (refer to diagram (d)) showing only a saddle stitch shading part, a right end shading part and a left end shading part is obtained. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to a technique for processing image data read by a reading device such as a scanner.

  When copying a saddle-stitched original such as a book or a magazine, it is necessary to set the original with the original face down on a reading table such as a copying machine or a scanner. Then, the image surface of the original is obtained by illuminating the original surface and detecting the reflected light by an optical sensor such as a CCD.

  At this time, if the document is set in a spread state, the saddle stitching portion is lifted from the reading table, so that a shaded portion is formed in the vicinity of the saddle stitching portion in the read image data, and the appearance is impaired. was there.

As a method of removing the shadow portion, focusing on the fact that the saddle stitching portion extends linearly over the entire length of the document with respect to the direction orthogonal to the spread direction of the document, unlike characters and figures. It is known that a line extending over the entire length is detected as a saddle stitch portion, and an image in a predetermined range is erased (masking) around that portion (see, for example, Patent Document 1).
JP-A-10-255024

  However, in the above method, when a relatively long line parallel to the saddle stitching portion exists in the vicinity of the saddle stitching portion, this long line may be detected as the saddle stitching portion. There is a problem that a part of the original image is deleted by mistake.

  SUMMARY An advantage of some aspects of the invention is that it provides an image processing method, an image processing apparatus, an image forming apparatus, and an image processing system capable of accurately detecting a shadow portion.

  The image processing method according to claim 1, which has been made to achieve such an object, includes a plurality of pixel data obtained by reading an image on a document and including light and dark parameters representing light and dark. From the image data that can form an image corresponding to the original image by arranging the pixels based on the two-dimensional arrangement, the light / dark abnormal portion in which the light / dark parameter does not reflect the original light / dark of the original image is extracted. An image processing method for performing orthogonal scanning orthogonal to the scanning direction from a light / dark integrated value obtained by integrating the values of the light / dark parameter for each of pixel data arrays arranged in a preset scanning direction in the image data Light / dark integral value distribution processing step for obtaining the light / dark integral value distribution in the direction, and spatial frequency analysis of the light / dark integral value distribution obtained by the light / dark integral value distribution processing step In order to extract the light and dark abnormal portion from the frequency analysis step to be performed and the frequency components in the low frequency region lower than the preset normal frequency region from the result of the spatial frequency analysis performed by the frequency analysis step A frequency component extraction step for extracting a frequency component to be processed based on a frequency component strength condition set in advance, and an inverse transform of the extracted frequency component extracted by the frequency component extraction step with respect to the spatial frequency analysis The extracted light / dark integral value distribution processing step for obtaining the extracted light / dark integral value distribution in the orthogonal scanning direction of the image data constituted by the extracted frequency components, and based on the extracted light / dark integral value distribution, And a light / dark abnormality portion determination step for determining the presence or absence of light.

  According to the image processing method configured as described above, the pixel data corresponding to the light / dark abnormality portion is obtained by reversing the spatial component of the frequency component in the low frequency region lower than the normal frequency region in the spatial frequency analysis result. Since the image can be extracted by performing conversion, a portion forming a wide image along the orthogonal scanning direction is a candidate for a light / dark abnormal portion. That is, in the pixel data, when there is a long line along the scanning direction, when the spatial frequency analysis is performed on the pixel data corresponding to the long line, the spatial frequency of the frequency component corresponding to the long line is Since it becomes sufficiently higher than the frequency component in the low frequency region, it is not extracted as a light / dark abnormal part.

Therefore, even when a long line exists in the vicinity of the light / dark abnormal part, the light / dark abnormal part can be detected correctly without erroneously detecting the long line as the light / dark abnormal part.
The normal frequency region may include a frequency component obtained by performing spatial frequency analysis on the light-dark integrated value distribution of the image data in the region where the long line along the scanning direction is formed. As described in item 2, the normal frequency region is set based on a frequency component obtained by performing spatial frequency analysis on a light-dark integrated value distribution for image data in a region where characters and figures are formed. Also good.

  According to the image processing method configured as described above, since the spatial frequency of the frequency component corresponding to the portion where the character or figure is formed is lower than the spatial frequency of the frequency component corresponding to the long line, the space in the normal frequency region Wide frequency range.

  For this reason, the spatial frequency range of the low frequency region used for determining the light / dark abnormal part is narrowed, and in the frequency component extraction step, the frequency component to be processed is extracted from the frequency components in the low frequency region. The processing load can be reduced.

  When the pixel data corresponding to the saddle-stitched portion when the double-page spread document is read is detected as at least a light / dark abnormality portion, the line of the saddle-stitched portion extends along the direction orthogonal to the double-page spread direction. Therefore, if the above-mentioned light-dark integral value distribution is obtained for the image data when the spread original is read and the spatial frequency analysis of this light-dark integral value distribution is performed, it corresponds to the saddle stitch portion. The intensity of the frequency component is often the strongest. Therefore, as described in claim 3, the frequency component extracting step may extract at least a frequency component having the strongest intensity in the low frequency region.

  In addition, when detecting pixel data corresponding to the saddle-stitched portion and both end portions when reading a double-page spread document as at least the light and dark abnormality portion, the lines of the saddle-stitched portion and both end portions are orthogonal to the spread direction of the spread original. Since the spread document exists continuously from one end to the other end of the spread document, the above-mentioned brightness integral value distribution is obtained for the image data when the spread document is read, and the spatial frequency analysis of the brightness integral value distribution is performed. In many cases, the frequency components corresponding to the saddle stitch portion and the both end portions are the top three strongest. For this reason, as described in claim 4, the frequency component extraction step may extract the top three frequency components having a strong intensity in the low frequency region.

  Further, in the image processing method according to claim 4, the width along the orthogonal scanning direction of the light and dark abnormal part at one end part in the both end parts and the orthogonal scanning direction of the light and dark abnormal part at the other end part of the both end parts. However, the spatial frequency of the frequency component corresponding to the light and dark abnormal part at one end and the frequency component corresponding to the light and dark abnormal part at the other end are different, but the light and dark at one end and the other end are different. When the width of the abnormal part along the orthogonal scanning direction is the same, the frequency components at one end and the other end have the same spatial frequency. The frequency components other than the portion are extracted. Therefore, as described in claim 5, the frequency component extracting step extracts a frequency component having an intensity equal to or higher than a predetermined value from among the top three frequency components having a high intensity. Is desirable.

  According to the image processing method configured as described above, the intensity of the frequency components other than the saddle stitching portion and both end portions is smaller than the intensity of the frequency component of the saddle stitching portion and both end portions. Even when the width of the abnormal portion along the orthogonal scanning direction is the same, it is possible to prevent pixel data other than the saddle stitching portion and the both end portions from being a light and dark abnormal portion.

  The saddle stitching portion is often located at the center of the image data in the scanning direction. Therefore, in the image processing method according to claim 3, as described in claim 6, the light / dark abnormality portion. The determination step determines whether or not there is a peak in a predetermined range that corresponds to a central portion in the scanning direction of the document image in the image data in the extracted light / dark integral value distribution, If the peak is within a predetermined range, the pixel data corresponding to the peak may be used as the light / dark abnormal portion data.

According to the image processing method configured as described above, it is possible to identify the light / dark abnormal portion of the saddle stitch portion.
In addition, since the saddle stitching portion is often located at an intermediate portion between both end portions, in the image processing method according to claim 4 or 5, as described in claim 7, the light / dark abnormality portion determination step. In the extracted light-dark integrated value distribution, whether or not there is a peak in the middle of both end peaks based on the positions of the both end peaks extracted from both end peaks indicating both ends of the original image. In some cases, the peak may be determined as the saddle stitch portion peak, and the pixel data corresponding to the peak may be used as the light / dark abnormality portion data.

According to the image processing method configured as described above, it is possible to identify the light / dark abnormal portion of the saddle stitch portion.
Further, in the image processing method according to claim 6 or 7, when the saddle stitch portion does not exist along the orthogonal scanning direction, that is, when the saddle stitch portion exists along the scan direction, Since there is no peak corresponding to the saddle stitch portion in the light / dark integrated value distribution, the light / dark abnormal portion cannot be extracted. For this reason, as described in claim 8, the image data rotation step of rotating the image data by 90 ° with respect to the scanning direction is included, and the light / dark abnormality portion determination step does not extract the light / dark abnormality portion data. In this case, it is desirable that the light / dark integral value distribution processing step is executed again after the image data rotation step is executed.

  According to the image processing method configured as described above, when the saddle stitch portion exists along the scanning direction due to a difference in the orientation in which the document is set, the light / dark abnormal portion data is extracted in the light / dark abnormal portion determination step. Therefore, the image data rotation step rotates the image data by 90 ° with respect to the scanning direction. As a result, the saddle stitch portion is present along the orthogonal scanning direction, so that the light / dark abnormal portion can be extracted.

  Further, in the image processing method according to any one of claims 1 to 8, as described in claim 9, the original image data is compared with the light / dark abnormality portion data extracted by the light / dark abnormality portion determination step. There may be provided a light / dark abnormality portion correcting step for performing correction so as to approach the light / dark state.

  According to the image processing method configured as described above, even when a long line exists in the vicinity of the light / dark abnormality portion, the long line is erroneously detected as the light / dark abnormality portion in the light / dark abnormality portion determination step. Therefore, it is possible to perform correction so that the light / dark abnormal portion approaches the original light / dark state of the image data without correcting the portion that is not the light / dark abnormal portion.

  In this case, the correction is performed so as to approach the original light and dark state of the image data. For example, as described in claim 10, the pixel data array arranged in the orthogonal scanning direction of the image data The spatial frequency analysis of the brightness / darkness parameter distribution with respect to the orthogonal scanning direction of the pixel data string is performed, and a frequency component corresponding to the brightness / darkness abnormal portion data extracted by the frequency component extraction step is obtained from the result of the spatial frequency analysis. Some of them are removed, and the result of the removed spatial frequency analysis is subjected to inverse transformation with respect to the spatial frequency analysis.

  According to the image processing method configured as described above, the frequency component corresponding to the portion forming the wide image along the orthogonal scanning direction is removed from the result of the spatial frequency analysis of the pixel data string. Even when a long line or character is included in the light / dark abnormal part, the light / dark abnormal part can be removed without erasing the long line or character.

  In the image processing method according to claim 9, as described in claim 11, for correction, the light / dark abnormal portion data is divided by the value of the total number of pixels in the scanning direction of the image data. Brightness / darkness abnormal portion data may be obtained, and the lightness / darkness parameter correction may be performed on each pixel data array arranged in the orthogonal scanning direction of the image data based on the correction lightness / darkness abnormal portion data.

  According to the image processing method configured in this way, it is not necessary to perform spatial frequency analysis and inverse transformation with respect to spatial frequency analysis, and it is only necessary to perform light and dark parameter correction based on the correction light and dark abnormal part data. Can be reduced.

  In addition, in claim 11, specifically, the light / dark parameter correction based on the correction light / dark abnormal portion data means that, as described in claim 12, the spatial frequency in the spatial frequency analysis result is 0. Based on the intensity of the frequency component, it is determined whether or not the background portion in the image data can be considered bright. If it is determined that the background portion can be considered bright, the orthogonality of the image data is determined. For each of the pixel data rows arranged in the scanning direction, the correction light / dark abnormality portion data is subtracted, and if it is determined that the background portion is not bright, the image data is arranged in the orthogonal scanning direction. The correction light / dark abnormality portion data is added to each of the pixel data trains.

  That is, the intensity of the frequency component having a spatial frequency of 0 in the spatial frequency analysis result reflects the value of the brightness parameter of the background portion, and the brightness of the background portion is determined based on this strength. If the background portion is considered bright, it is determined that the light / dark abnormal portion is darker than the background portion, and correction is performed so as to approach the original light / dark state by subtracting the correction light / dark abnormal portion data. On the other hand, if the background part is not bright, it is determined that the light and dark abnormal part is brighter than the background part, and correction is made to approximate the original light and dark state by adding correction light and dark abnormal part data. doing.

  Furthermore, in the image processing method according to any one of claims 9 to 12, when the image data is monochrome image data, the light / dark abnormality is corrected by the light / dark abnormality portion correcting step as described in claim 13. It is determined whether or not the background portion of the image data that has been corrected for the partial data can be regarded as white, and if it is determined that the background portion can be considered as white, the pixel data constituting the background portion The density value of pixel data having a density equal to or lower than a first predetermined value that can be regarded as white that is close to the minimum density value that can be expressed in the black and white image data is converted to the minimum density value, and the background portion cannot be regarded as white. If it is determined that the density value of the pixel data having a density equal to or higher than a second predetermined value that can be regarded as black that is close to the maximum density value that can be expressed in the black and white image data among the pixel data constituting the background portion. Maximum darkness It may have a first background conversion step of converting the value.

  According to the image processing method configured as described above, a pixel having a density equal to or lower than the first predetermined value that can be regarded as white close to the minimum density value is white as the minimum density value, and black is regarded as close to the maximum density value. A pixel having a density equal to or higher than the second predetermined value is black with the maximum density value. That is, when there is a density change from the minimum density value to the first predetermined value in the background portion, the white background surface is unified to the minimum density value, and there is a density change from the second predetermined value to the maximum density value in the background portion. In this case, since the black background of the maximum density value is unified, it is possible to generate image data constituting an easy-to-see image with no density change in the background.

  Furthermore, in the image processing method according to any one of claims 9 to 12, when the image data is color image data, the light / dark abnormality is corrected by the light / dark abnormality portion correcting step as described in claim 14. The hue of the background portion in the image data corrected for the partial data is detected, and the pixel data having a hue within a predetermined range that can be regarded as a hue close to the detected hue among the pixel data constituting the background portion. A second background conversion step for converting the hue into the detected hue may be included.

  According to the image processing method configured as described above, a pixel having a hue within a predetermined range that can be regarded as a hue close to the detected hue among the pixels constituting the background portion becomes the detected hue. In other words, when there is a hue change within a predetermined range centered on the detected hue in the background portion, the detected hue is unified, so that it is possible to generate image data that constitutes an easy-to-view image with no hue change in the background portion. .

  Further, in the image processing method according to any one of claims 9 to 12 and claim 14, when the image data is color image data, the light / dark abnormality portion correcting step as described in claim 15 is performed. The image data that has been corrected for the light / dark abnormal portion data may be converted into a specified color specification format.

  According to the image processing method configured as described above, since the image data is converted into the color specification format designated according to the output destination of the image data, it is possible to save the trouble of converting the color specification format at the output destination of the image data. Can do.

  In addition, when the image data is monochrome image data, the brightness of the pixel data is usually expressed by density. Therefore, in the image processing method according to any one of claims 1 to 13, the image data is monochrome image data. In this case, as described in claim 16, it is desirable that the brightness parameter is a density.

According to the image processing method configured as described above, it is possible to save the trouble of converting the density of pixel data into the light and dark parameters used by the image processing method.
In addition, when the image data is color image data, the brightness of the pixel data is usually represented by brightness, and therefore the image processing according to any one of claims 1 to 12 and claims 14 to 15. In the method, when the image data is color image data, it is preferable that the brightness parameter is brightness as described in claim 17.

According to the image processing method configured as described above, it is possible to save the trouble of converting the brightness of the pixel data into the brightness parameter used by the image processing method.
An image processing apparatus according to claim 18 is obtained by reading an image on a document, and is composed of a plurality of pixel data including a light and dark parameter representing light and dark, and pixels based on the pixel data are arranged in a two-dimensional manner. Thus, an image processing apparatus for extracting from the image data that can form an image according to the original image, an abnormal light / dark part in which the light / dark parameter does not reflect the original light / dark of the original image. From the light / dark integral value obtained by integrating the value of the light / dark parameter for each of the storage means for storing the image data and the pixel data array arranged in the preset scanning direction in the image data, A light / dark integral value distribution processing means for obtaining a light / dark integral value distribution in the orthogonal scanning direction, and a spatial frequency solution of the light / dark integral value distribution obtained by the light / dark integral value distribution processing means. And a frequency analysis means for performing the above-mentioned, and from the result of the spatial frequency analysis performed by the frequency analysis means, a preset is set in order to extract the light / dark abnormal part in a low frequency region lower than a preset normal frequency region. The frequency component extraction means for extracting the frequency component to be processed based on the intensity condition, and the extraction frequency component extracted by the frequency component extraction means by performing an inverse transform on the spatial frequency analysis, thereby extracting the frequency component Extracted light / dark integrated value distribution processing means for obtaining an extracted light / dark integrated value distribution in the orthogonal scanning direction of image data composed of frequency components, and determining whether there is a light / dark abnormal portion based on the extracted light / dark integrated value distribution Part determining means.

The image processing apparatus configured as described above realizes the image processing method according to claim 1, and can obtain the same effects as the image processing method according to claim 1.
In the image processing apparatus according to claim 18, when the storage means does not have a capacity capable of storing the entire image data, a part of the image on the original is read and the read image data is stored. After storing in the means, the light-dark integrated value distribution is obtained for the stored image data, and then the remaining portion of the image on the original is read again, and the read image data is stored in the storage means, and then the stored image data For example, it is possible to read the image on the document little by little and sequentially obtain the light / dark integral value distribution as in the case of obtaining the light / dark integral value distribution. Therefore, the process becomes complicated. Therefore, as described in claim 19, it is desirable that the storage means has a capacity capable of storing the entire image data.

  According to the image processing apparatus configured as described above, after reading the entire image on the document and storing the entire image data in the storage unit, the light / dark integral value distribution processing unit calculates the light / dark integral value distribution of the entire image data. Therefore, it is possible to simplify the process of reading the document and the process of obtaining the light / dark integral value distribution without repeating the process of reading the document and the process of obtaining the light / dark integral value distribution.

  An image forming apparatus according to claim 20 is an image forming apparatus that reads an image printed on a recording medium and forms an image based on the read image data. The image processing apparatus described above is provided.

According to the image forming apparatus configured as described above, an effect similar to that of the image processing apparatus according to claim 18 can be obtained.
An image processing system according to a twenty-first aspect includes an image processing device according to the eighteenth or nineteenth aspect, a reading unit that reads an image printed on a recording medium and acquires image data, and an image An image forming means for forming an image based on the data, and the reading means and the image processing apparatus are connected to each other so that the image data acquired by the reading means can be transmitted to the image processing apparatus. The image forming unit and the image processing device are connected so that image data in which the light / dark abnormality portion is corrected by the light / dark abnormality portion correcting unit included in the image processing apparatus can be transmitted to the image forming unit. To do.

  According to the image processing system configured as described above, an effect similar to that of the image processing device according to claim 18 can be obtained. Furthermore, the reading unit and the image forming unit can be arranged in different places within a range that can be connected to the image processing apparatus, and the degree of freedom of arrangement is improved.

(Embodiment 1)
Hereinafter, an image processing system configured using an image processing apparatus will be described as a first embodiment of the present invention.

First, the configuration of the image processing system 1 will be described with reference to FIG.
The image processing system 1 includes a scanner 10 that reads a document as image data, an inkjet printer 20 that forms a color image on recording paper based on the image data, and an image processing apparatus 100 that processes the image data.

  As shown in FIG. 1, the image processing apparatus 100 includes a CPU 102 that executes processing based on a predetermined processing program, a hard disk (hereinafter referred to as HDD) 104 as a storage device, a ROM 106 that stores various control programs, A RAM 108 for storing data input from an external device, an input / output interface (hereinafter referred to as an input / output I / F) 110 for inputting image data from the scanner 10 and outputting image data to the inkjet printer 20, and a user operation A user interface (hereinafter referred to as a user I / F) 112 including a plurality of possible operation keys 112a and a display panel 112b for displaying various types of information is provided, and is configured by a known personal computer.

The HDD 104 has a storage area for an image processing program 104a for the CPU 102 to execute image processing.
The RAM 108 has a capacity capable of storing the entire input image data D1, which will be described later.

  The user I / F 112 outputs a copy start command to the CPU 102 when an operation for starting copying is performed by the operation key 112a. Further, an operation for designating whether the scanner 10 reads a document as a black and white or color image, or designating whether to set a high image quality mode in which image data to be output to the ink jet printer 20 is set to a high image quality. Can be performed via the operation key 112a.

  In the image processing apparatus 100 configured as described above, first, the CPU 102 captures image data input from the scanner 10 and stores it in the RAM 108. Thereafter, the CPU 102 reads image data from the RAM 108 and performs image processing based on the image processing program 104a. Then, the image data subjected to image processing is output to the inkjet printer 20 via the input / output I / F 110.

Next, the outline of the image processing of the present invention will be described with reference to FIG.
FIG. 2A shows input image data D <b> 1 read from the scanner 10. The input image data D1 is obtained by setting a saddle-stitched book or the like on the reading table of the scanner 10 in a spread state and reading it with the scanner 10.

  In the input image data D1, X_SIZE pixel data is arranged in the horizontal direction (hereinafter also referred to as X-axis direction), and Y_SIZE pixel data is arranged in the vertical direction (hereinafter also referred to as Y-axis direction) (X_SIZE: input data horizontal). Size (pixel): Y_SIZE: Input data vertical size (pixel)). That is, it is composed of X_SIZE × Y_SIZE pixel data P (X, Y). X is the horizontal coordinate value of the pixel, Y is the vertical coordinate value, 0, 1, 2,..., (X_SIZE-1) for X, 0, 1, 2,. ..., the value of (Y_SIZE-1) is given. X_SIZE and Y_SIZE differ depending on the image data. Further, the pixel data P (X, Y) has information that 0 is black and 255 is white in the range of 0 to 255, and this information has density values in the range of 0 to 100 based on a table or function. It has been converted. Usually, 0 is the brightest at the minimum density value, and 100 is the darkest at the maximum density value.

  In the input image data D1, a saddle stitch portion D1d that exists as a long line in the Y-axis direction at the center, a saddle stitch shadow portion D1a that is darker than the periphery in the vicinity of the saddle stitch portion D1d, and darker than the periphery at the right end There are a right end shadow portion D1b, a left end shadow portion D1c darker than the surroundings at the left end, and a plurality of characters and figures.

  Then, in the input image data D1, density integrated values obtained by integrating density values for one line of pixel data strings along the Y-axis direction are obtained for all X coordinate values from 0 to (X_SIZE-1). Note that the direction in which density values for one line of pixel data string are integrated is set in advance as the Y-axis direction. That is, the Y-axis direction is the scanning direction in the present invention, the X-axis direction is the orthogonal scanning direction in the present invention, and the scanning direction is set for a spread original such as a book regardless of the reading direction of the scanner 10. The direction parallel to the saddle stitch portion D1d needs to be set as the scanning direction.

The histogram H1 shown in FIG. 2B shows the X coordinate of the pixel data string on the horizontal axis and the density integrated value of the pixel data string on the vertical axis.
In this histogram H1, there are a peak H1a in the vicinity of X = (X_SIZE / 2), a peak H1b in the vicinity of X = (X_SIZE-1), a peak H1c in the vicinity of X = 0, and a peak H1a, a peak H1b, and a peak H1c are These correspond to the saddle stitch shadow portion D1a, the right end shadow portion D1b, and the left end shadow portion D1c, respectively.

  Further, when Fourier transform is performed on the data represented by the histogram H1, data represented by the spatial frequency distribution F1 as shown in FIG. 2C is obtained. The peak F1a in the spatial frequency distribution F1 corresponds to the peak H1a of the histogram H1, and the peak F1b corresponds to the peak H1b and the peak H1c. That is, the peak H1a in the histogram H1 rises more slowly than the peaks H1b and H1c, so that the peak F1a has a lower spatial frequency than the peak F1b in the spatial frequency distribution F1. Moreover, since the rise of the peak H1b and the peak H1c is substantially the same, it appears as a peak F1b having the same spatial frequency.

  Then, the peak F1a and the peak F1b are strong peaks at a spatial frequency (for example, 1 cycle / mm or less) lower than the frequency component obtained by Fourier transforming the density integral value of the image data in the region where the characters are formed. When the inverse Fourier transform is performed on the data represented by the extracted peak, only the saddle stitched shadow portion D1a, the right end shadow portion D1b, and the left end shadow portion D1c as shown in FIG. Data represented by a histogram H2 indicating the density integrated value is obtained. Incidentally, unlike FIG. 2A, in the input image data D1, when the saddle stitch portion D1d is orthogonal to the Y-axis direction, the peak H1a corresponding to the saddle stitch shadow portion D1a cannot be extracted. . In this case, the input image data D1 is rotated by 90 ° in order to make the saddle stitch portion D1d and the Y-axis direction parallel to each other. Thereby, the Y-axis direction becomes the scanning direction in the present invention.

Then, after the data represented by the histogram H2 is obtained, the processing described below is performed for each of the pixel data strings arranged in the X-axis direction of the input image data D1.
That is, the pixel data string on the scanning line SC1 shown in FIG. 2A will be described as an example. As shown in FIG. 3A, the vicinity of the saddle stitched shadow portion D1a on the scanning line SC1 is “G , A character D1f representing “H”, a character D1g representing “E”, and a character D1h representing “F” are arranged, and the horizontal axis of the pixel on the scanning line SC1 of this portion is represented by a pixel When a histogram having the X coordinate of the data P (X, Y) and the density on the vertical axis is created, a histogram H3 as shown in FIG. 3B is created.

  In this histogram H3, peak H3a, peak H3d, peak H3e, peak H3f, peak H3g, and peak H3h correspond to the saddle stitched portion D1a, saddle stitch portion D1d, character D1e, character D1f, character D1g, and character D1h, respectively. Yes. Although not clearly shown in FIG. 3A, the density of the saddle stitching portion D1d gradually increases from the periphery of the saddle stitching portion D1d toward the center, and therefore corresponds to the saddle stitching portion D1d. The peak H3d has a shape in which the concentration gradually increases.

  Further, when Fourier transform is performed on the data represented by the histogram H3, data represented by the spatial frequency distribution F2 as shown in FIG. 3C is obtained. In FIG. 3C, the peak F2a having the lowest spatial frequency is the peak H3a, the peak F2b having the second lowest spatial frequency is the peak H3d, and the peak F2c having the highest spatial frequency is the peak H3e, the peak H3f, and the peak H3g. And peak H3h.

  That is, since the peak H3a has the slowest rise in the histogram H3, the peak F2a has the lowest spatial frequency in the spatial frequency distribution F2. Furthermore, since the peak H3d rises more slowly than the peaks H3e, H3f, H3g, and H3h, the peak F2b has the second lowest spatial frequency in the spatial frequency distribution F2. The rises of the peak H3e, the peak H3f, the peak H3g, and the peak H3h are substantially the same, and thus appear as a peak F2c having the same spatial frequency.

  Since the spatial frequency of the peak F2a of the spatial frequency distribution F2 matches the spatial frequency of the peak F1a of the spatial frequency distribution F1, in the data represented by the spatial frequency distribution F2, the same space as that extracted by the spatial frequency distribution F1 is used. When the frequency component data is removed, the data corresponding to the peak F2a is removed.

  Further, when inverse Fourier transform is performed on the data represented by the spatial frequency distribution F2 from which the peak F2a has been removed, as shown in FIG. 3D, the histogram H4 from which the peak H3a has been removed from the histogram H3. Is obtained.

Then, when the above-described processing is performed on all the pixel data strings arranged in the X-axis direction, input image data D1 from which the shaded portion D1a has been removed is obtained as shown in FIG.
Next, image processing executed by the CPU 102 of the image processing apparatus 100 will be described with reference to FIG. FIG. 5 is a flowchart showing image processing. In the following description, the same words and phrases as described in the above outline are used in the same meaning as the words in the outline.

  When this image processing is executed, the CPU 102 first determines whether or not a copy start command is input via the operation key 112a in S10. If it is determined that a copy start command has been input (S10: YES), the process proceeds to S20. On the other hand, if it is determined that the copy start command has not been input (S10: NO), the process of S10 is repeated.

  When the process proceeds to S20, an original set in the scanner 10 is read by the scanner 10 before the copy start command is input, and input image data D1 (see FIG. 2A) is generated. In S30, the shadow / light detection flag provided in the RAM 108 is cleared. In the following description, setting a flag indicates that the value of the flag is 1, and clearing the flag indicates that the value of the flag is 0. Further, in S40, the read input image data D1 is stored in the RAM 108.

  Next, in S50, it is determined whether or not the input image data D1 is a monochrome image. Here, before the copy start command is input, whether or not the original is read as a monochrome image or a color image is specified via the operation key 112a. (S50: YES), the process proceeds to S70. On the other hand, if it is designated to read as a color image, it is determined that the image is not a monochrome image (S50: NO), and the process proceeds to S60.

  In S60, the data input image D1 stored in the RAM 108 is converted into a well-known L * C * h color specification format. Note that L represents lightness, C represents saturation, and h represents hue. For example, regarding the lightness, three types of color components defined by R (red), G (green), and B (blue) of 256 gradations of 0 to 255, respectively, included in each pixel data P (X, Y). Is converted into a brightness value in the range of 0 to 100 by a table or function. Then, the process proceeds to S70.

  In S70, when the data input image D1 is a black and white image, density integrated values obtained by integrating density values for one line of pixel data strings along the Y-axis direction are set to all X coordinates of 0 to (X_SIZE-1). The pixel data string is obtained. The data obtained here is represented by a histogram H1 shown in FIG. On the other hand, if the data input image D1 is not a black and white image, all X coordinate pixels from 0 to (X_SIZE-1) are integrated with the brightness values obtained by integrating the brightness values of one line of pixel data along the Y-axis direction. Calculate for data columns. As shown in FIG. 4A, the data obtained here is represented by a histogram H1L in which the horizontal axis is the X coordinate of the pixel data P (X, Y) and the vertical axis is the brightness integrated value. The brightness has a value in the range of 0 to 100. In contrast to the density, 0 is usually the darkest and 100 is the brightest. That is, the histogram H1L is obtained by inverting the integrated density value with respect to the histogram H1 shown in FIG.

  In S80, Fourier transform is performed on the data obtained in S70, thereby obtaining data represented by a spatial frequency distribution as shown in FIG. The histogram H1 and the histogram H1L have the same magnitude of change in the value of the vertical axis with respect to the X-axis direction, although there is a difference that the value of the vertical axis of the histogram is inverted. Therefore, when the Fourier transform is performed on the data represented by the histogram H1L, the first term F1Lc, which is a component having a spatial frequency of 0, is added to the spatial frequency distribution F1, as shown in FIG. 4B. Data represented by the spatial frequency distribution F1L is obtained.

  Next, in S90, similarly to S50, it is determined whether or not the input image data D1 is a monochrome image. If it is determined that the input image data D1 is a monochrome image (S90: YES), the process proceeds to S100. On the other hand, if it is determined that the input image data D1 is not a monochrome image (S90: NO), the process proceeds to S110.

  In S100, it is determined whether or not the intensity of the first term in the spatial frequency distribution F1 is large. This first term is a component having a spatial frequency of 0 in the spatial frequency distribution F1, and corresponds to the background density of the data input image D1. That is, the strength of the first term increases as the background density increases. In FIG. 2A, since the background density value of the input image data D1 is 0, the value of the first term of the spatial frequency distribution F1 is 0.

  Then, when the intensity of the first term is, for example, 30% or more with respect to the value obtained by integrating the intensities of all the spatial frequency components of the spatial frequency distribution F1, it is determined that the intensity of the first term is large (S100: YES). ), The process proceeds to S130. On the other hand, if the strength of the first term is less than 30%, for example, it is determined that the strength of the first term is small (S100: NO), and the process proceeds to S120. That is, when the strength of the first term is 30% or more, the background is considered black, and when it is less than 30%, the background is considered white.

  In S110, it is determined whether the intensity of the first term in the spatial frequency distribution F1L is small. This first term is a component corresponding to the lightness of the background of the data input image D1. That is, the intensity of the first term increases as the lightness of the background increases.

  Then, when the intensity of the first term is less than 70%, for example, with respect to the value obtained by integrating the intensities of all the spatial frequency components of the spatial frequency distribution F1L, it is determined that the intensity of the first term is small (S110: YES). ), The process proceeds to S120. On the other hand, if the strength of the first term is 70% or more, for example, it is determined that the strength of the first term is high (S110: NO), and the process proceeds to S130. That is, the background is dark when the intensity of the first term is less than 70%, and the background is bright when it is 70% or more.

  When the process proceeds to S120, a shadow part detection process is performed. That is, the shadow portion detection process is executed when the background is considered white. This shadow portion detection process is executed according to the procedure shown in FIG. FIG. 6 is a flowchart showing the shadow detection process.

  When this shadow portion detection process is executed, first, in S310, the CPU 102, in the spatial frequency distribution represented by the data obtained in S70, for example, with a spatial frequency of 1 cycle / mm or less, the total spatial frequency of the spatial frequency distribution. If the average value of the intensity in the component is AVG and the maximum intensity in the spatial frequency distribution is MAX, it is determined whether there is a frequency component having an intensity of “[(MAX−AVG) / 2} + AVG]” or more. That is, the determination is made to extract only the peak F1a and the peak F1b in the spatial frequency distribution F1, or to extract only the peak F1La and the peak F1Lb in the spatial frequency distribution F1L.

  If it is determined that there is a frequency component having an intensity equal to or higher than “[(MAX−AVG) / 2} + AVG]” at a spatial frequency of 1 cycle / mm or less (S310: YES), the process proceeds to S320. Data corresponding to a frequency component that satisfies the condition is extracted. That is, only data corresponding to the peak F1a and the peak F1b in the spatial frequency distribution F1 is extracted, or only data corresponding to the peak F1La and the peak F1Lb is extracted in the spatial frequency distribution F1L.

  Then, in S330, by performing inverse Fourier transform on the data corresponding to the peak extracted in S320, the saddle stitch shadow part D1a, the right end shadow part D1b, and the left end shadow part as shown in FIG. Data represented by a histogram H2 having peaks H2a, H2b, and H2c corresponding to each of D1c is obtained.

  Further, in S340, peak positions P2a, P2b, and P2c, which are X coordinate values having the strongest intensity in each of the peaks H2a, H2b, and H2c, are obtained for the histogram H2 (see FIG. 2D). Then, it is determined whether or not the peak position P2a is within a predetermined range (for example, (P2b + P2c) / 2 ± 10 mm) centered on the value of (P2b + P2c) / 2. Here, when it is determined that the value of the peak position P2a is within the predetermined range (S340: YES), the process proceeds to S350, it is determined that “there is a shadow part”, and the shadow part detection process is ended. On the other hand, when it is determined that the peak position P2a is not within the predetermined range (S340: NO), the process proceeds to S360, where it is determined that “there is no shadow part”, and the shadow part detection process ends.

  That is, in the case of a saddle-stitched document such as a book, an intermediate position between the position of the right end portion and the position of the left end portion is a saddle stitch portion, so that the peak H2a is in the vicinity of the center position obtained from the peaks H2b and H2c. Has determined that the peak H2a is a shadow portion caused by the saddle stitching portion. Similarly, when it is not in the vicinity of the center position, it is determined that the shadow portion is not caused by the saddle stitch portion.

  Returning to S310, if it is determined that there is no frequency component having an intensity of [(MAX−AVG) / 2} + AVG] at a spatial frequency of 1 cycle / mm or less (S310: NO), the process proceeds to S360. It is determined that there is no shadow part, and the shadow part detection process is terminated.

  Returning to FIG. 5, when the process proceeds to S <b> 130, the light detection unit detection process is performed. That is, the scale portion detection process is executed when the background is regarded as black. This lever part detection process is executed according to the procedure shown in FIG. FIG. 7 is a flowchart showing the light detection unit detection process.

  “Tekari” means that when a document with a dark background is set in a scanner with the page facing apart, the saddle stitching part is lifted from the reading table. This means that reflection occurs on the surface of the skin and the background becomes white.

  After executing this lever part detection process, the CPU 102 first determines in S410 whether or not there is a frequency component having an intensity of [(MAX−AVG) / 2} + AVG] or more, as in the process of S310. To do.

  For example, if it is determined that there is a frequency component having an intensity of [(MAX−AVG) / 2} + AVG] at a spatial frequency of 1 cycle / millimeter or less (S410: YES), the process proceeds to S420, and the process of S320 is performed. Similarly, frequency components satisfying the conditions in S410 are extracted.

  Then, in S430, by performing inverse Fourier transform on the data corresponding to the peak extracted in S420, the saddle stitch shadow part D1a, the right end shadow part D1b, and the left end shadow part as shown in FIG. Data represented by a histogram H2 having peaks H2a, H2b, and H2c corresponding to each of D1c is obtained.

  Further, in S440, as in the process of S340, peak positions P2a, P2b, and P2c, which are X coordinate values having the strongest intensity in each of the peaks H2a, H2b, and H2c, are obtained for the histogram H2 (FIG. 2 (d)). Then, it is determined whether or not P2a is located within a predetermined range (for example, (P2b + P2c) / 2 ± 10 mm) centered on the value of (P2b + P2c) / 2. Here, if it is determined that the value of P2a is located within the predetermined range (S440: YES), the process proceeds to S350, it is determined that “there is a lighted portion”, and the lighted portion detection process ends. On the other hand, if it is determined that the value of P2a is not located within the predetermined range (S440: NO), the process proceeds to S460, where it is determined that there is no “lighting portion”, and the lightening portion detection process is terminated.

  Returning to S410, if it is determined that there is no frequency component having an intensity of [(MAX−AVG) / 2} + AVG] at a spatial frequency of 1 cycle / mm or less (S410: NO), the process proceeds to S460, and “ It is determined that there is no “lighting portion”, and the light detection portion detection process ends.

  Returning to FIG. 5, when the shadow portion detection process in S120 is completed, it is determined in S140 whether or not there is a shadow portion. That is, when it is determined that there is a “shadow part” in S120 (see S350), it is determined that there is a shadow part (S140: YES), and the process proceeds to S160. On the other hand, if it is determined in S120 that there is no shadow (see S360), it is determined that there is no shadow (S140: NO), and the process proceeds to S170.

  Similarly, when the focus part detection process in S130 is completed, it is determined in S150 whether or not there is a focus part. That is, if it is determined in S130 that there is a “lighting portion” (see S450), it is determined that there is a lighting portion (S140: YES), and the process proceeds to S160. On the other hand, if it is determined in S130 that there is no “lighting part” (see S460), it is determined that there is no lighting part (S150: NO), and the process proceeds to S170.

If the process proceeds to S160, correction processing is performed. This correction processing is executed according to the procedure shown in FIG. FIG. 8 is a flowchart showing the correction process.
When this correction process is executed, the CPU 102 first performs a Fourier transform correction process in S510. This Fourier transform correction process is executed according to the procedure shown in FIG. FIG. 9 is a flowchart showing the Fourier transform correction process.

When this Fourier transform correction process is executed, the CPU 102 first sets Y to 0 as the initial setting of the Y coordinate value of the pixel data P (X, Y) in S710.
Then, the process proceeds to S720, and Fourier transform is performed on the data representing the X-direction distribution (see FIG. 3B) of the density (brightness) value of the pixel data string having the Y coordinate value indicated by Y. Data represented by a spatial frequency distribution as shown in FIG.

  Further, in S730, data corresponding to the spatial frequency component having the same spatial frequency as the spatial frequency component extracted in S320 or S420 is removed from the data obtained in S720.

  Next, in S740, a histogram obtained by removing the peak H3a corresponding to the saddle stitch portion D1a as shown in FIG. 3D by performing inverse Fourier transform on the data processed in S340. The data represented is obtained. In S750, the data obtained in S740 is reflected in the density (lightness) value of the pixel data string having the Y coordinate value indicated by Y in the data input image D1, and the process proceeds to S760.

  In step S760, Y is incremented. Thereafter, the process proceeds to S770, where it is determined whether or not the value of Y is equal to or greater than Y_SIZE. If it is determined that the value of Y is smaller than Y_SIZE (S770: NO), the process proceeds to S720 and the above process is repeated. On the other hand, if it is determined that the value of Y is greater than or equal to Y_SIZE (S770: YES), the Fourier transform correction process is terminated.

  Returning to the correction process of FIG. 8, when the Fourier transform correction process of S <b> 510 is completed, the process proceeds to S <b> 520 to determine whether or not the high-quality mode is set. The high image quality mode is set by operating the operation key 112a by the operator before starting copying. If it is determined that the high image quality mode is set (S520: YES), the process proceeds to S530. On the other hand, if it is determined that the high image quality mode is not set (S520: NO), the correction process is terminated.

  When the process proceeds to S530, it is determined whether or not the input image data D1 is a monochrome image in the same manner as S50. If it is determined that the input image data D1 is a black and white image (S530: YES), the process proceeds to S540. On the other hand, if it is determined that the input image data D1 is not a monochrome image (S530: NO), the process proceeds to S570.

  When the process proceeds to S540, it is determined whether or not a shadow portion has been detected in the image processing of FIG. If it is determined that a shadow portion has been detected (S540: YES), the process proceeds to S550, and the pixel data P (X, Y having a density value of 10% or less of the maximum density value in the image data corrected in S510, for example. ) Is set to 0, and the correction process is terminated. On the other hand, if it is determined that the shadow portion has not been detected, it is determined that the focus portion has been detected (S540: NO), and in the image data corrected in S510, for example, a density value of 90% or more of the maximum density value is obtained. The density value of the pixel data P (X, Y) it has is set to 100, and the correction process is terminated.

  When the process proceeds to S570, the number of pixels having the same (L * C * h) is obtained for all the pixels of the image data corrected in S510. In S580, (L * C * h) having the largest number of pixels is extracted, and h of the extracted (L * C * h) is determined as the hue of the background.

  Further, in S590, the hue of the pixel data P (X, Y) having a hue angle within a predetermined range (for example, hue angle ± 3 °) around the hue determined in S580 is determined in S580. The correction process is terminated after conversion to a new hue.

Returning to FIG. 5, when the correction process of S160 is completed, the process proceeds to S200.
When the process proceeds to S170, it is determined whether or not the shadow / light detection flag is cleared. If it is determined that the shadow / light detection flag is not cleared (S170: NO), the process proceeds to S200. On the other hand, if it is determined that the shadow / light detection flag is cleared (S170: YES), the process proceeds to S180, the shadow / light detection flag is set, and further stored in the RAM 108 in S190. The data input image D1 being rotated is rotated by 90 °. And it transfers to S70 and repeats said process.

  When the process proceeds to S200, it is determined whether or not the input image data D1 is a monochrome image in the same manner as S50. If it is determined that the input image data D1 is a monochrome image (S200: YES), the process proceeds to S220. On the other hand, if it is determined that the input image data D1 is not a monochrome image (S200: NO), the input image data D1 is converted from the L * C * h color specification format to the well-known RGB color specification format in S210. R, G, and B represent red, green, and blue, respectively.

In step S220, the input image data D1 stored in the RAM 108 is output to the inkjet printer 20, and the image processing ends.
In the embodiment described above, the processing of S70 in FIG. 5 is the bright / dark integrated value distribution processing step in the present invention, the processing of S80 in FIG. 5 is the frequency analysis step in the present invention, and the processing in S320 in FIG. The extraction step, S330 in FIG. 6 and S430 in FIG. 7 are the extracted light / dark integral value distribution processing step in the present invention, S340 to S360 in FIG. 6 and S440 to S460 in FIG. It is.

  5 is the image data rotation step in the present invention, the process in S160 in FIG. 5 is the light / dark abnormality correction step in the present invention, and the processes in S540 to S560 in FIG. 8 are the first background conversion step in the present invention. 8, the process of S570 to S590 is a second background conversion step in the present invention, and the process of S210 in FIG. 5 is a format conversion step in the present invention.

The RAM 108 is a storage unit in the present invention, the scanner 10 is a reading unit in the present invention, and the inkjet printer 20 is an image forming unit in the present invention.
The “shaded part” and the “lighting part” are the light and dark abnormal parts in the present invention, the Y axis direction is the scanning direction in the present invention, the X axis direction is the orthogonal scanning direction in the present invention, and the density integral value and the lightness integral value are the present In the present invention, the light / dark integral value, histogram H1 and histogram H1L are the light / dark integral value distribution in the present invention, Fourier transform is the spatial frequency analysis in the present invention, inverse Fourier transform is the inverse transform to the spatial frequency analysis in the present invention, and peak H2a is in the present invention. The extracted light / dark integral value distribution and histogram H2 are light / dark abnormal portion data in the present invention.

  Further, “10% of the maximum density value” in S550 is the first predetermined value in the present invention, “90% of the maximum density value” in S550 is the second predetermined value in the present invention, and the predetermined range in S590 is the present invention. Is a predetermined range.

  According to the image processing apparatus 100 of the present embodiment configured as described above, a density integrated value obtained by integrating the density value for each of the pixel data strings arranged in the Y-axis direction in the input image data D1 is obtained (S70). The Fourier transform is performed on the data represented by the histogram H1 (S80). Furthermore, from the data represented by the spatial frequency distribution F1, a spatial frequency lower than a frequency component obtained by Fourier-transforming the density integral value of the image data in the region where the characters are formed, and a predetermined intensity or more. A peak is extracted (S320), and the histogram H2 is obtained by performing inverse Fourier transform on the data represented by the extracted peak (S330). Then, in the histogram H2, peaks H2b and H2c indicating both end portions of the original image are extracted, and it is determined whether or not the peak H2a is at an intermediate portion between the peaks H2b and H2c. And if it determines that it exists in an intermediate part, it will determine peak H2a as a shadow part (light part) (S340, S350 (S440, S450)).

  That is, the pixel data corresponding to the shadow portion (light portion) can be extracted by performing an inverse Fourier transform on a frequency component having a low spatial frequency set in advance in the spatial frequency distribution F1. A portion that forms a wide image along the X-axis direction is a candidate for a shadow portion (lighting portion). That is, in the pixel data, when there is a long line along the Y-axis direction, when the Fourier transform is performed on the pixel data corresponding to the long line, the spatial frequency of the frequency component corresponding to the long line is Since it is sufficiently higher than the frequency component in the low frequency region, it is not extracted as a shadow portion (lighting portion).

  Therefore, even if a long line exists in the vicinity of the shadow part (lighting part), the shadow part (lighting part) is correctly detected without erroneously detecting the long line as a shadow part (lighting part). Can be detected.

  Further, the image processing apparatus 100 performs a Fourier transform of the density distribution with respect to the Y-axis direction for each of the pixel data strings arranged in the X-axis direction of the input image data D1 (S720), and further peaks from the result of the Fourier transform. The frequency component corresponding to H2a is removed (S730), and the result of the removed Fourier transform is inversely transformed with respect to the Fourier transform (S740), so that a long line or character is included in the shaded portion (light portion). Even in such a case, it is possible to remove the shadow portion (light portion) without erasing long lines or characters.

  When the image processing apparatus 100 determines that “there is no shadow part (lighting part)” (S360 (S460)), the input image data D1 is rotated by 90 °. Portion) can be extracted along the X-axis direction.

  Further, the image processing apparatus 100 sets the density value of the pixel data P (X, Y) having a density value of 10% or less of the maximum density value to 0 (S550), and has a density value of 90% or more of the maximum density value. Since the density value of the pixel data P (X, Y) is set to 100, it is possible to generate image data that forms an easy-to-view image with no density change in the background portion.

  Further, the image processing apparatus 100 determines the hue of the pixel data P (X, Y) having a hue angle within a predetermined range (for example, hue angle ± 3 °) with the hue determined in S580 as the center in S580. Since the hue is converted to the determined hue (S590), it is possible to generate image data that forms an easy-to-view image with no hue change in the background portion.

  Since the RAM 108 has a capacity capable of storing the entire input image data D1, the entire image on the original is read and the entire image data is stored in the RAM 108, and then the process of S70 is performed. it can.

(Embodiment 2)
Hereinafter, a configuration in which the configuration of the image forming apparatus of the present invention is applied to a copying apparatus will be described as a second embodiment of the present invention.

First, the configuration of the copy apparatus 200 will be described with reference to FIG.
As shown in FIG. 10, the copying apparatus 200 includes a CPU 202 that executes processing based on a predetermined processing program, a ROM 206 that stores various control programs, a RAM 208 that stores data input from an external device, and the like. A user interface 212 comprising a plurality of operable keys 212a and a display panel 212b for displaying various information, a scanner 214 for reading a document as image data, a scanner interface for data communication with the scanner 214 (hereinafter referred to as scanner I) 216, a printer engine 218 that forms a color image on recording paper based on image data, and a printer engine interface (hereinafter referred to as printer engine I / F) 220 that performs data communication with the printer engine 218. It has.

In the ROM 206, a storage area for an image processing program 206a for the CPU 202 to execute image processing is secured.
The printer engine 220 includes Y, M, C, and K ink nozzles, and includes a print head capable of printing with large, medium, and small dots, and ink cartridges of each of Y, M, C, and K colors. Ink can be ejected downward to form an image on the print medium.

  In the first embodiment, the CPU 102 executes image processing (see FIG. 5), but in the second embodiment of the present invention, the CPU 202 in the copying apparatus 200 executes. Since the image processing method is the same as that of the first embodiment, description thereof is omitted.

According to the copying apparatus 200 configured as described above, an image can be formed using image data generated in the same manner as in the first embodiment.
For this reason, even when a long line exists in the vicinity of the shadow portion (light portion), the long line is correctly detected without being erroneously detected as the shadow portion (light portion). Can be detected.

As mentioned above, although one Example of this invention was described, this invention is not limited to the said Example, A various aspect can be taken.
For example, in the present embodiment, an example in which a configuration as an image forming apparatus is applied to the copy apparatus 200 is illustrated. However, as long as it has a function of inputting and outputting image data, it can be similarly applied to a printer, a facsimile, and the like.

  In the present embodiment, the image data obtained by reading the image on the document is also image-processed. However, the same applies to the case where the image data obtained by photographing using a camera or the like is image-processed. It is applicable to.

  In the present embodiment, in the processes of S340 and S440, the peak H2a is a shadow portion caused by the saddle stitch portion based on the positions of the peaks H2b and H2c corresponding to the right end shadow portion D1b and the left end shadow portion D1c, respectively. It was determined whether or not. However, it is determined whether or not the peak H2a is located in the vicinity of the center (X = X_SIZE / 2) in the X-axis direction of the input image data D1, and when the peak H2a is located in the vicinity, the peak H2a is at the saddle stitch portion. It may be determined that the shadow portion is caused.

In the present embodiment, the input image data D1 is rotated by 90 ° in the process of S190. However, the scanning direction may be rotated by 90 °.
Further, in the present embodiment, the processing configured by the procedure of FIG. 3 is shown as the processing for correcting the image data. However, it may be configured by the procedure shown in FIG.

In other words, the pixel column on the scanning line SC1 shown in FIG. 2A will be described as an example. As shown in FIG. 11A, the vicinity of the saddle stitched portion D1a on the scanning line SC1 is “G”. A character D1e representing “H”, a character D1f representing “H”, a character D1g representing “E”, and a character D1h representing “F” are arranged, and the horizontal axis of the pixel on this scanning line SC1 is represented by pixel data. When a histogram with the X coordinate of P (X, Y) and the density on the vertical axis is created, a histogram H3 as shown in FIG. 11B is created (FIG. 11A and FIG. 11B are respectively shown). (It is the same as FIG. 3 (a) and FIG. 3 (b).)
Then, with respect to the data represented by the histogram H2 shown in FIG. 2D, the value obtained by dividing the integrated density value by the number of pixels Y_SIZE in the Y-axis direction is obtained, so that the saddle stitched shadow part D1a and the right edge shadow part D1b are obtained. , The correction data represented by the histogram H5 indicating the density distribution for one pixel of only the left end shadow portion D1c is obtained (FIG. 11C shows the distribution in the vicinity of the saddle stitched shadow portion D1a. The corresponding peak H5a is shown). The correction data represented by the histogram H5 is correction light / dark abnormality portion data in the present invention.

  Furthermore, for each X coordinate value, when the density value represented by the histogram H5 is subtracted from the density value represented by the histogram H3, the peak H3a is removed from the histogram H3 as shown in FIG. Data represented by the histogram H4 is obtained. At this time, portions of peak H3d, peak H3f, and peak H3g that overlap with peak H3a are simultaneously removed, so that the concentration values of peak H3d, peak H3f, and peak H3g become small as shown in FIG. .

  Then, when the above-described processing is performed on all the pixel data strings arranged in the Y-axis direction, as shown in FIG. 11E, the shadow portion D1a is removed and the character portion that overlaps the shadow portion D1a Is obtained.

The process shown in FIG. 11 can be realized by executing a simple correction process described below instead of the Fourier transform correction process of S510 in the correction process of FIG.
Next, simple correction processing executed by the CPU 102 of the image processing apparatus 100 will be described with reference to FIG. FIG. 12 is a flowchart showing the simple correction process.

  When this simple correction process is executed, the CPU 102 first sets Y to 0 as the initial setting of the Y coordinate value in S910. Then, the process proceeds to S920, and for the data obtained by the inverse Fourier transform in S330 or S430, the integral value is divided by the number of pixels Y_SIZE in the Y-axis direction, so that one pixel as shown in FIG. Correction data representing the density (lightness) distribution of the.

  Then, the process proceeds to S930, and similarly to S90, it is determined whether or not the input image data D1 is a monochrome image. If it is determined that the input image data D1 is a black and white image (S930: YES), the process proceeds to S940. On the other hand, if it is determined that the input image data D1 is not a monochrome image (S930: NO), the process proceeds to S950.

  In S940, similarly to S100, it is determined whether or not the intensity of the first term in the spatial frequency distribution F1 is large. Then, when the intensity of the first term is, for example, 30% or more with respect to the value obtained by integrating the intensities of all the spatial frequency components of the spatial frequency distribution F1, it is determined that the intensity of the first term is large (S940: YES) ), The process proceeds to S1000. On the other hand, when the strength of the first term is less than 30%, for example, it is determined that the strength of the first term is small (S940: NO), and the process proceeds to S960. The first term is a frequency component having a spatial frequency of 0 in the present invention.

  In S950, similarly to S110, it is determined whether or not the intensity of the first term in the spatial frequency distribution F1L is small. Then, when the intensity of the first term is less than 70%, for example, with respect to the value obtained by integrating the intensities of all spatial frequency components of the spatial frequency distribution F1L, it is determined that the intensity of the first term is small (S950: YES) ), The process proceeds to S960. On the other hand, if the strength of the first term is 70% or more, for example, it is determined that the strength of the first term is large (S950: NO), and the process proceeds to S1000.

That is, it is assumed that the background is bright when the process proceeds to S960, and the background is dark when the process proceeds to S1000.
When the process proceeds to S960, correction data obtained in S920 is obtained from the density (lightness) value of each pixel data P (X, Y) in the pixel data string having the Y coordinate value indicated by Y. The density (lightness) value corresponding to the X coordinate value is subtracted. That is, the correction data is corrected as a shaded part. Thereafter, in S970, the data processed in S960 is reflected in the density (lightness) value of the pixel data string having the Y coordinate value indicated by Y of the input image D1 data.

  Further, in S980, Y is incremented. Thereafter, the process proceeds to S990, where it is determined whether or not the value of Y is equal to or greater than Y_SIZE. If it is determined that the value of Y is smaller than Y_SIZE (S990: NO), the process proceeds to S960 and the above process is repeated. On the other hand, if it is determined that the value of Y is greater than or equal to Y_SIZE (S990: YES), the simple correction process is terminated.

  When the process proceeds to S1000, correction data obtained in S920 is obtained from the density (lightness) value of each pixel data P (X, Y) in the pixel data string having the Y coordinate value indicated by Y. The density (lightness) value corresponding to the X coordinate value at is added. That is, the correction data is corrected as a measuring portion. Thereafter, in S1010, the data processed in S1000 is reflected in the density (lightness) value of the pixel data string having the Y coordinate value indicated by Y of the input image D1 data.

  Further, in S1020, Y is incremented. Thereafter, the process proceeds to S1030, and it is determined whether or not the value of Y is equal to or greater than Y_SIZE. If it is determined that the value of Y is smaller than Y_SIZE (S1030: NO), the process proceeds to S1000 and the above process is repeated. On the other hand, if it is determined that the value of Y is greater than or equal to Y_SIZE (S1030: YES), the simple correction process is terminated.

  According to the image processing apparatus 100 of the present embodiment configured as described above, correction data obtained by dividing the integral value by the number of pixels Y_SIZE in the Y-axis direction is obtained for the data obtained by the inverse Fourier transform in S330 or S430. Determination (S920), based on the intensity of the first term in the spatial frequency distributions F1 and F1L, it is determined whether or not the background portion in the input image data D1 can be considered bright (S930 to S950). If it is determined that the background portion is bright, the correction data is assumed to correspond to the shaded portion, and the correction data is subtracted for each of the pixel data strings arranged in the X-axis direction ( S960-S990). On the other hand, if it is determined that the background portion is not bright, the correction data is added to each of the pixel data strings arranged in the X-axis direction, assuming that the correction data corresponds to the measuring portion. (S960-S990).

  That is, it is not necessary to perform Fourier transform and inverse Fourier transform, and it is sufficient to perform correction based on the correction data, so that the processing load can be reduced. However, the brightness and darkness of the pixel data corresponding to the long lines and characters included in the shaded part (lighting part) is out of the original light and dark state due to the influence of the correction data.

1 is a block diagram illustrating a configuration of an image processing apparatus. Outline | summary explanatory drawing of a shadow part detection. Outline of image correction Explanatory drawing of the histogram and spatial frequency distribution which used the brightness. The flowchart showing an image processing procedure. The flowchart showing the shadow part detection processing procedure. The flowchart showing the procedure part detection process procedure. 10 is a flowchart showing a correction processing procedure. The flowchart showing the Fourier-transform correction process procedure. FIG. 2 is a block diagram illustrating a configuration of a copy apparatus. FIG. 6 is a schematic explanatory diagram of image correction processing according to another embodiment. The flowchart showing the simple correction process procedure.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Image processing system, 10 ... Scanner, 20 ... Inkjet printer, 100 ... Image processing apparatus, 102 ... CPU, 104 ... HDD, 106 ... ROM, 108 ... RAM, 110 ... Input / output I / F, 112 ... User I / F, 112a ... operation keys, 112b ... display panel, 200 ... copy device, 202 ... CPU, 206 ... ROM, 208 ... RAM, 212 ... user I / F, 212a ... operation keys, 212b ... display panel, 214 ... scanner, 216 ... Scanner I / F, 218 ... Printer engine, 220 ... Printer engine I / F.

Claims (21)

An image corresponding to the original image is formed by two-dimensionally arranging the pixels based on the pixel data, which is obtained by reading an image on the original and is composed of a plurality of pixel data including light and dark parameters representing light and dark. An image processing method for extracting, from possible image data, a light / dark abnormal portion in which the light / dark parameter does not reflect the original light / dark of the document image,
A light / dark integrated value distribution in an orthogonal scanning direction orthogonal to the scanning direction is obtained from a light / dark integrated value obtained by integrating the value of the light / dark parameter for each of pixel data sequences arranged in a preset scanning direction in the image data. An integrated value distribution processing step;
A frequency analysis step for performing a spatial frequency analysis of the light-dark integral value distribution obtained by the light-dark integral value distribution processing step;
From the result of the spatial frequency analysis performed by the frequency analysis step, a frequency set in advance for extracting the light / dark abnormal portion from the frequency components in the low frequency region lower than the preset normal frequency region. A frequency component extraction step for extracting a frequency component to be processed based on a component intensity condition;
Extracting the extracted frequency component extracted in the orthogonal scanning direction of the image data constituted by the extracted frequency component by performing inverse transformation on the spatial frequency analysis with respect to the extracted frequency component extracted in the frequency component extracting step A light-dark integrated value distribution processing step;
Based on the extracted light / dark integrated value distribution, a light / dark abnormal portion determination step for determining the presence / absence of a light / dark abnormal portion;
An image processing method comprising: The normal frequency region is set based on a frequency component obtained by performing spatial frequency analysis on a light-dark integrated value distribution for image data in a region where characters and figures are formed. Image processing method. Detecting at least the pixel data corresponding to the saddle-stitched portion at the time of reading a double-page spread as the light and dark abnormality portion,
The image processing method according to claim 1, wherein the frequency component extraction step extracts at least a frequency component having the strongest intensity in the low frequency region. Detecting at least the pixel data corresponding to the saddle stitching portion and both end portions when reading a double-page spread as the light / dark abnormality portion,
The image data is obtained by reading a spread original including both ends in the spread direction and the saddle stitched saddle stitch portion,
The frequency component extraction step extracts the top three frequency components having a strong intensity in the low frequency region.
The image processing method according to claim 3. 5. The image processing method according to claim 4, wherein the frequency component extraction step extracts a frequency component having an intensity equal to or higher than a predetermined value from among the top three frequency components having high intensity. In the light / dark abnormal portion determination step, in the extracted light / dark integral value distribution, it is determined whether or not there is a peak in a predetermined range corresponding to a central portion in the scanning direction of the document image in the image data. 4. The image processing method according to claim 3, wherein when the peak is within the predetermined range, the pixel data corresponding to the peak is used as the light / dark abnormality portion data. In the extracted light / dark integral value distribution, the light / dark abnormal portion determination step extracts both end peaks indicating both end portions of the document image, and based on the positions of the both end peaks, 5. The method according to claim 4, wherein it is determined whether or not there is a peak, in which case the peak is determined as the saddle stitch portion peak, and pixel data corresponding to the peak is used as the light-dark abnormality portion data. The image processing method according to claim 5. An image data rotation step of rotating the image data by 90 ° with respect to the scanning direction;
In the light / dark abnormality portion determination step, when the light / dark abnormality portion data is not extracted, the light / dark integral value distribution processing step is executed again after executing the image data rotation step,
The image processing method according to claim 6 or 7, wherein: 9. The light / dark abnormal part correction step of correcting the light / dark abnormal part data extracted in the light / dark abnormal part determination step so as to approach the original light / dark state of the image data. The image processing method according to any one of the above. The light / dark abnormality portion correcting step performs the spatial frequency analysis of the light / dark parameter distribution with respect to the orthogonal scanning direction of the pixel data string for each of the pixel data strings arranged in the orthogonal scanning direction of the image data, and The frequency component corresponding to the light / dark abnormal part data extracted in the frequency component extraction step is removed from the result of the spatial frequency analysis, and the result of the removed spatial frequency analysis is inversely transformed with respect to the spatial frequency analysis. The image processing method according to claim 9. The light / dark abnormal portion correction step obtains light / dark abnormal portion data for correction obtained by dividing the light / dark abnormal portion data by the value of the total number of pixels in the scanning direction of the image data, and the orthogonal scanning direction of the image data. 10. The image processing method according to claim 9, wherein the brightness / darkness parameter correction is performed on each of the pixel data strings arranged in a manner based on the correction light / dark abnormality portion data. The light / dark abnormality portion correcting step determines whether or not the background portion in the image data can be considered bright based on the intensity of the frequency component having a spatial frequency of 0 in the spatial frequency analysis result, If it is determined that it is bright, the correction light / dark abnormal portion data is subtracted for each pixel data sequence arranged in the orthogonal scanning direction of the image data, and the background portion is considered bright. If it is determined that there is no correction, for each of the pixel data sequences arranged in the orthogonal scanning direction of the image data, the correction light / dark abnormality portion data is added,
The image processing method according to claim 11. The image data is black and white image data,
When it is determined whether the background portion in the image data corrected for the light / dark abnormality portion data in the light / dark abnormality portion correction step can be regarded as white, and when it is determined that the background portion can be regarded as white The density value of pixel data having a density equal to or lower than a first predetermined value that can be regarded as white that is close to the minimum density value that can be expressed in the black-and-white image data is converted into the minimum density value. When it is determined that the background portion cannot be regarded as white, the pixel data constituting the background portion is equal to or greater than a second predetermined value that can be regarded as black that is close to the maximum density value that can be expressed in the monochrome image data. A first background conversion step of converting the density value of the pixel data of the density to the maximum density value;
The image processing method according to claim 9, wherein the image processing method is performed. The image data is color image data,
The hue of the background portion in the image data that has been corrected for the light / dark abnormal portion data in the light / dark abnormal portion correction step is detected, and can be regarded as a hue close to the detected hue in the pixel data constituting the background portion. A second background conversion step of converting the hue of pixel data having a hue within a predetermined range into the detected hue;
The image processing method according to claim 9, wherein the image processing method is performed. The image data is color image data,
The method according to claim 9, further comprising a format conversion step of converting the image data that has been corrected for the light / dark abnormal portion data in the light / dark abnormal portion correcting step into a specified color specification format. The image processing method according to claim 14. The image data is black and white image data,
The brightness parameter is density.
The image processing method according to any one of claims 1 to 13, wherein the image processing method is performed. The image data is color image data,
The brightness parameter is brightness.
The image processing method according to any one of claims 1 to 12 and claims 14 to 15. An image corresponding to the original image is formed by two-dimensionally arranging the pixels based on the pixel data, which is obtained by reading an image on the original and is composed of a plurality of pixel data including light and dark parameters representing light and dark. An image processing apparatus for extracting from the possible image data, a light / dark abnormal part in which the light / dark parameter does not reflect the original light / dark of the document image,
Storage means for storing the image data;
A light / dark integrated value distribution in an orthogonal scanning direction orthogonal to the scanning direction is obtained from a light / dark integrated value obtained by integrating the value of the light / dark parameter for each of pixel data sequences arranged in a preset scanning direction in the image data. Integral value distribution processing means;
Frequency analysis means for performing spatial frequency analysis of the light-dark integral value distribution obtained by the light-dark integral value distribution processing means;
From the result of the spatial frequency analysis performed by the frequency analysis means, in a low frequency region lower than a preset normal frequency region, based on a strength condition set in advance to extract the light and dark abnormal part, A frequency component extracting means for extracting a frequency component to be processed;
Extracting the extracted frequency component extracted by the frequency component extracting means to obtain the extracted light-dark integral value distribution in the orthogonal scanning direction of the image data constituted by the extracted frequency component by performing an inverse transform on the spatial frequency analysis A light-dark integrated value distribution processing means;
Based on the extracted light / dark integrated value distribution, a light / dark abnormal part determining means for determining the presence / absence of a light / dark abnormal part;
An image processing apparatus comprising: The image processing apparatus according to claim 18, wherein the storage unit has a capacity capable of storing the entire image data. An image forming apparatus that reads an image printed on a recording medium and forms an image based on the read image data,
An image forming apparatus comprising the image processing apparatus according to claim 18. An image processing device according to claim 18 or 19,
Reading means for reading an image printed on a recording medium and acquiring image data;
Image forming means for forming an image based on the image data;
With
The reading means and the image processing apparatus are connected so that the image data acquired by the reading means can be transmitted to the image processing apparatus, and
The image forming means and the image processing apparatus are connected so that the image data in which the light / dark abnormality portion is corrected by the light / dark abnormality portion correcting means provided in the image processing apparatus can be transmitted to the image forming means.
An image processing system characterized by that.

Images