JP5822750B2 - Apparatus and method for detecting impurities in read image - Google Patents

Description

  The present invention relates to an impurity detection apparatus and method, and more particularly to an apparatus and method for detecting impurities, particularly impurities contained in a document sheet, with high accuracy when reading a document image.

  With the spread of printers in emerging markets where the income of individuals and corporations is low, the output paper handled by users is increasing in quality compared to the past. As a result, in an image reading apparatus such as a document reader, the quality of the paper composing the original to be read tends to decrease.

  Also, in existing mature markets, as a part of efforts to reduce running costs and environmental burdens by individuals and corporations, low-quality paper with slightly lower quality such as recycled paper is used instead of high-quality paper. Is becoming more common. On the other hand, in order to promote environmental efforts, some efforts have been made to handle paper with the taste of recycled paper by suppressing the use of paper pulp bleach used in papermaking. .

  Compared with high-quality paper, these low-quality papers are characterized by a decrease in whiteness, which indicates the average brightness of the entire paper, and an increase in the appearance frequency of impurities contained in the paper, also called dirt. Have. As a result, there is a demand for higher-performance impurity detection technology for scanned images. In particular, the impurities in the paper tend to be higher in concentration than the impurities deposited on or attached to the image reading apparatus. For this reason, there is a need for a more accurate detection technique that not only extracts from the background of the paper but also can be distinguished from printed or small objects such as small characters.

  On the other hand, many detection techniques aiming at reduction / removal of impurities from a read image have been proposed and implemented.

  For example, there is a technique for referring to image data as a material for determination, calculating coordinate information, and detecting dust (for example, see Patent Document 1). In addition, there is a technique for classifying whether an object is a scratch or dust according to the size of the outline of the object in the image (see, for example, Patent Document 2).

Japanese Patent Laid-Open No. 9-83805 JP-A-6-50905

  However, in Patent Document 1 and Patent Document 2, it is not possible to distinguish between the impurities in the paper and a symbol with a small number of points (for example, a decimal point in arithmetic numbers, a period / comma in English). As a result, in the worst case, there is a problem that the impurity cannot be detected, or a character at a small point is erroneously detected as an impurity and erased.

  Furthermore, in Patent Document 2, objects are classified using labeling, and the sizes of the respective contours are calculated. However, when such a determination method is performed, since the processing is heavy, the circuit scale and There is also a cost problem that leads to an increase in memory.

Impurity detecting apparatus according to the present invention, the pixel of interest in the object in the input image, of the first pixel group composed of the peripheral pixels including the adjacent pixels of the target pixel, the pixel that is part of the object The number of pixels that are part of the object in the second pixel group that includes a means for counting the number as a first count value and a peripheral pixel that includes pixels adjacent to the outside of the first pixel group and means for counting a second count value, of the third group of pixels consisting of peripheral pixels including a pixel adjacent to the outer side of the second pixel group, the number of pixels that are part of the object means for counting a third count value, the first count value and the second count value and on the basis of the third count value, the pixel of interest is an impurity der in the input image Or a a determining means, said determining means, at least, the first count value does not exceed the first threshold value, and the third count value does not exceed the second threshold value, and , if the value based on the sum with at least the said third count value and second count value does not exceed the third threshold value, the pixel of interest is characterized that you determined to be impurities.

  According to the present invention, it is possible to realize high-precision impurity detection with low-load processing using only the read image data without requiring any special operation or hardware change in the image reading apparatus itself. Become. This is particularly effective when detecting impurities in the scanned image due to impurities contained in the paper. It is possible to distinguish between impurities and small-point characters and symbols to prevent false detection. It becomes.

It is a block diagram which shows the structure of the impurity detection apparatus in one Embodiment of this invention. FIG. 6 is a diagram illustrating an operation flow of the first embodiment. It is a figure which shows an input image and an object signal. It is a figure which shows the arrangement pattern of each pixel group with respect to an attention pixel. It is a figure which shows an input image and an object signal. It is a figure which shows the count value in each pixel. It is a figure which shows an impurity flag. It is a figure which shows the result image of the process for impurity removal. It is a figure which shows the arrangement pattern of each pixel group with respect to an attention pixel. It is a figure which shows an object signal. It is a figure which shows the arrangement pattern of each pixel group with respect to an attention pixel. FIG. 10 is a diagram illustrating an operation flow of the second embodiment. It is a figure which shows the arrangement pattern of each pixel group with respect to an attention pixel. It is a figure which shows the order of an image process.

  The best mode for carrying out the present invention will be described below with reference to the drawings. However, the components described in this embodiment are merely examples, and the scope of this description is not intended to be limited thereto.

  In the present embodiment, an apparatus for impurity detection according to the present invention will be described using a document reader having a function of reading an image typified by a scan function. However, the present invention provides an apparatus having a function including image reading such as a copy function and a FAX function in addition to a scanning function (that is, an image reading apparatus such as an electrophotographic multifunction peripheral or an inkjet multifunction peripheral). ) Is also applicable. Further, the present invention does not have an image reading function in the apparatus itself, and performs predetermined image processing on the read image transmitted through the connection after being electrically connected to the image reading apparatus. The present invention can also be applied to such a device. That is, for example, the present invention can be applied to an image processing apparatus such as an image up-converter.

  FIG. 1 is a block diagram illustrating a configuration of an impurity detection device according to an embodiment. Reference numeral 100 denotes an impurity detection device in the present embodiment. Reference numeral 111 denotes a main control unit that controls the entire impurity detection apparatus. Reference numeral 112 denotes a storage unit that stores image data and information associated therewith. Reference numeral 113 denotes an operation unit that receives commands from the operator and displays information to the operator. Reference numeral 114 denotes an impurity removing unit that removes impurities from the input image. Reference numeral 101 denotes an image input unit for inputting a read image of a document as electronic image data. Reference numeral 102 denotes an object detection unit that detects whether each pixel in the image data is a part of an object described later. Reference numeral 103 denotes a characteristic acquisition unit that acquires characteristics relating to the object of each pixel. Reference numeral 104 denotes a comprehensive determination unit that comprehensively determines whether each pixel is an impurity.

  With reference to FIG. 2, the flow of processing by the impurity detection apparatus in the present embodiment will be described. This process is executed under the control of the main control unit 111 according to the control program read from the storage unit 112.

  In step 201, the image input unit 101 inputs image data to be read according to an operation performed by the operator via the operation unit 113. The input image is stored in the storage unit 112.

  In step 202, the object detection unit 102 detects an object from the input image under the control of the main control unit 111. The object here is typified by text, graphics, images, etc., and is a generic name for anything printed or described, or something that was accidentally mixed in, except for the white background in the image. Impurity images are also included. Further, in this step, the object detection unit 102 determines whether each pixel of the input image is a part of the object. As a result of the determination, if the pixel is a part of the object, the object signal that is the detection result is generated as ON, and if it is not a part of the object, the object signal is generated as OFF.

  FIG. 3 shows an example of the object signal. In general, there are various methods for detecting an object from the input image 301. For example, a binarized signal 302 obtained from an input image using a fixed threshold value or a variation threshold value, an edge extraction signal 303 obtained by primary differentiation or secondary differentiation of the input image, or an image indicating characters obtained by image area separation There is a method of detecting an object by the area signal 304 or the like. The object may be detected by any one of these methods, or may be detected by another method.

  In the binarized signal 302, white pixels (for example, a pixel that is darker than the threshold in the input signal 301 becomes a white pixel in an inverted binarized image like the binarized signal 302). The object signal is ON, and the black pixel indicates that the object signal is OFF. As for the edge extraction signal 303 and the image area signal 304, the object signals of the pixels belonging to the object detected using these signals are set to ON, and the object signals of other pixels are set to OFF.

  Unless otherwise specified with respect to the subsequent figures, when an object signal is indicated, a white pixel is indicated as ON and a black pixel is indicated as OFF. Further, here, binary signals are taken up for easy understanding, but these object signals may be held as multi-level signals. The object signal generated for each pixel is stored in the storage unit 112 in association with the pixel.

  Steps 203 to 242 are characteristic points in this embodiment, and will be described in detail below.

  The processing from step 203 to step 242 is performed while shifting the target pixel one pixel at a time with respect to all the pixels in the input image. Hereinafter, a pixel in which these processes are performed (one pixel shifts when transitioning from step 243 to step 203) is referred to as a target pixel. Pixels around the target pixel to which the signal value is referred for processing of the target pixel are referred to as reference pixels. The reference pixel includes the target pixel itself.

  In the present embodiment, the reference pixels are divided into three groups (hereinafter referred to as pixel groups). These pixel groups are referred to as a first pixel group, a second pixel group, and a third pixel group, respectively. The arrangement of each pixel group with respect to the target pixel is illustrated in FIG. Each of the patterns 401 to 404 represents the arrangement of each pixel group with respect to the target pixel 421. These pixel groups become a first pixel group 411, a second pixel group 412, and a third pixel group 413 from areas close to the target pixel, respectively. The handling of these target pixels and pixel groups will be described later.

  The arrangement pattern of the pixel group may be any one of these patterns 401 to 404, and is not limited to these patterns. However, the arrangement pattern of the pixel group is configured as follows. In other words, the first pixel group includes the target pixel and its neighboring pixels. The second pixel group is composed of pixels adjacent to the outside of the first pixel group. The third pixel group is composed of pixels adjacent to the outside of the second pixel group. The first to third three pixel groups are arranged in layers in order with the target pixel as the center. It is important for the three pixel groups to be configured and arranged in this way in order to effectively bring out the effects of the present embodiment.

  In addition, the width of each layer of the first to third pixel groups is set so that impurities can be effectively detected by a process described later. The widths of the layers shown in the patterns 401 to 404 in this embodiment are set based on experiments so that impurities can be effectively detected. However, the width of each layer is not limited to this example, and can be freely set based on the size of the impurity to be detected and the degree of isolation of the impurity to be detected from other objects.

  In step 203, the main control unit 111 reads out the object signal associated with the target pixel from the storage unit 112, and checks whether it is ON. That is, it is confirmed whether or not the target pixel is a pixel in the object in the input image. If it is ON, the process transitions to step 204. If it is OFF, the processing moves to processing of the next pixel in the image.

  In step 204, the main control unit 111 resets the count values Cnt1, Cnt2, and Cnt3 used in the processing described later to zero. In other words, each count value is reset to 0 each time a new loop of pixels of interest is entered (that is, every time the process transitions from step 243 to step 203). Processing then transitions to processing block 251.

  The processing block 251 includes steps 211 to 214. In this processing block, the characteristic acquisition unit 103 performs a process of calculating the first count value Cnt1 under the control of the main control unit 111. Here, Cnt1 is a count value of a reference pixel having an ON object signal among reference pixels existing in the first pixel group (that is, a reference pixel that is a part of the object).

  First, in step 211, the characteristic acquisition unit 103 checks whether the object signal of the reference pixel of the first pixel group is ON. If it is ON, the process transitions to step 212. If it is OFF, the process transitions to the next reference pixel in the first pixel group.

In step 212, 1 is added to the first count value Cnt1.
If it is determined in step 213 that the processing has been completed for all the reference pixels in the first pixel group, the processing transitions to step 214. If there is a reference image that has not been processed, the process of step 211 is performed on the next reference image.

  In step 214, the first count value Cnt1 is associated with the target pixel and stored in the storage unit 112. Processing then transitions to processing block 252.

  The processing block 252 includes steps 221 to 224. In this processing block, the characteristic acquisition unit 103 performs a process of calculating the second count value Cnt2 under the control of the main control unit 111. Here, Cnt2 is a count value of reference pixels having an ON object signal among reference pixels existing in the second pixel group (that is, reference pixels that are part of the object).

  First, in step 221, the characteristic acquisition unit 103 checks whether or not the object signal of the reference pixel in the second pixel group is ON. If it is ON, the process transitions to step 222. If it is OFF, the process transitions to the next reference pixel in the second pixel group.

In step 222, 1 is added to the second count value Cnt2.
If it is determined in step 223 that the processing has been completed for all reference pixels in the second pixel group, the processing transitions to step 224. If there is a reference image that has not been processed, the process of step 221 is performed on the next reference image.

  In step 224, the second count value Cnt2 is associated with the target pixel and stored in the storage unit 112. Processing then transitions to processing block 253.

  The processing block 253 includes steps 231 to 234. In this processing block, the characteristic acquisition unit 103 performs a process of calculating the third count value Cnt3 under the control of the main control unit 111. Here, Cnt3 is a count value of reference pixels having an ON object signal among reference pixels existing in the third pixel group (that is, reference pixels that are part of the object).

  First, in step 231, the characteristic acquisition unit 103 checks whether the object signal of the reference pixel of the third pixel group is ON. If it is ON, the process transitions to step 232. If it is OFF, the process transitions to the next reference pixel in the third pixel group.

In step 232, 1 is added to the third count value Cnt3.
If it is determined in step 233 that the process has been completed for all reference pixels in the third pixel group, the process transitions to step 234. If there is a reference image that has not been processed, the process of step 231 is performed on the next reference image.

  In step 234, the third count value Cnt3 is associated with the target pixel and stored in the storage unit 112. Processing then transitions to processing block 254.

  The processing block 254 includes step 241 and step 242. In this processing block, the comprehensive determination unit 104 performs comprehensive determination using the count values Cnt1 to Cnt3 under the control of the main control unit 111. That is, in this processing block, it is determined whether or not the target pixel is a part of the impurity.

  First, in step 241, the comprehensive determination unit 104 reads three count values Cnt1, Cnt2, and Cnt3 from the storage unit 112 and performs comprehensive determination. Comprehensive determination is performed using the following conditional expressions (Formula 1 to Formula 3).

  Here, Th1, Th2, and Th3 are all predetermined constants. These constants are set as threshold values used for determination on the count value and the sum of the count values. Th1 is a threshold value related to the size of the object to which the target pixel belongs. Th2 and Th3 are thresholds related to the isolation degree of the object to which the target pixel belongs. When all of Equations 1 to 3 are satisfied, TRUE is returned as a result of the comprehensive determination. In the case of TRUE, it is determined that the target pixel is a part of the impurity, and the process transitions to Step 242. If there is a conditional expression that does not satisfy any one of Expressions 1 to 3, FALSE is returned as a result of comprehensive determination. In the case of FALSE, it is determined that the target pixel is not a part of the impurity, and the processing shifts to a processing loop of the next target pixel in the input image.

  That is, in step 241, the pixel of interest belongs in consideration of the number of pixels (first to third count values) that are part of the object in each of the first, second, and third pixel groups. The size and isolation of the object are determined. Based on this determination result, it is determined whether or not the target pixel is part of the impurity. Details of the determination of the size and isolation of the object to which the pixel of interest belongs will be described later.

  In step 242, the impurity flag is stored in the storage unit 112 in association with the target pixel. The impurity flag is a flag used in a process for removing impurities described later, and is a signal indicating an attribute associated with each pixel. An impurity flag that is ON indicates that the pixel is part of the impurity.

  When the processing of step 203 to step 242 is completed for all the pixels in the input image, in step 243, the main control unit 111 determines that there are no unprocessed pixels, and transitions to processing end.

  By performing the next image processing (not shown) using the generated impurity flag, the effect of removing impurities can be obtained. That is, the impurity removal unit 114 is a process of painting a pixel whose impurity flag is ON in white, or a process of making impurities inconspicuous by smoothing or the like.

  As described above, according to this embodiment, impurities are detected using the number of pixels that are part of an object in each of the first, second, and third pixel groups. By performing such detection, it is possible to distinguish impurities from small-point characters and symbols (for example, part of “i”, dot “.”), And prevent erroneous detection of impurities.

  Details of the significance of counting the number of pixels that are part of the object in the first, second, and third pixel groups will be described later.

The above is the configuration of the present embodiment related to the impurity detection apparatus. Hereinafter, specific examples of processing according to this configuration will be described using specific examples of processing signals. In this example, the pattern 401 is used as the arrangement pattern of the reference pixel group. The following values are used for the constants Th1, Th2, and Th3, respectively.
Th1 = 60
Th2 = 13
Th3 = 18

  Further, images 501 to 504 shown in FIG. 5 are used as images input to the image input unit 101 in step 201. The image 501 is an image representing the impurity 521 contained in the original sheet and its periphery. The image 502 is an image representing the decimal point 522 and its surroundings in the small point text. The image 503 is an image representing a part 523 of the alphabet “i” in the text having a number of points larger than the decimal point 522 and the periphery thereof. An image 504 is an image representing a halftone dot 524 of a photograph in the newspaper original and its periphery.

  Reference numerals 511 to 514 in FIG. 5 indicate the object signals generated by the object detection unit 102 in step 202.

  FIG. 6 shows the count values Cnt1, Cnt2 and Cnt3 of each pixel of the object corresponding to the reference numerals 521 to 524 in the areas 531 to 534 of FIG. That is, reference numerals 611 to 614 indicate the count value Cnt1 for each pixel counted in the processing block 251. Reference numerals 621 to 624 indicate the count value Cnt2 for each pixel counted in the processing block 252. Reference numerals 631 to 634 denote count values Cnt3 for the respective pixels counted in the processing block 253. For example, for the upper left pixel of the object 521, the reference numeral 611 shows the count value Cnt1 as 36, the reference numeral 621 shows the count value Cnt2 as 12, and the reference numeral 631 shows the count value Cnt3 as 0.

  FIG. 7 shows a result of the comprehensive determination unit 104 performing a comprehensive determination using Formulas 1 to 3 in the processing block 254 based on the three count values Cnt1, Cnt2, and Cnt3 (that is, an impurity flag). . A white portion of reference numeral 701 indicates that the impurity flag of each pixel of the object corresponding to reference numeral 531 is ON. Reference numerals 702, 703, and 704 indicate that the impurity flag of each pixel of the object corresponding to each of reference numerals 532, 533, and 534 is OFF.

  Using the generated impurity flag, the above-described process for removing impurities is performed. FIG. 8 shows an example of the effect. Reference numerals 801 to 804 are generated images when a process of painting the input image of the pixel in white is performed when the impurity flag of each pixel is turned ON. Reference numerals 811 to 814 denote image area signals when processing is performed in which the image area signal of the pixel is canceled and turned OFF when the impurity flag of each pixel is turned on.

  As described above, the count values Cnt1, Cnt2, and Cnt3 counted by the characteristic acquisition unit 103 are not only mere object signal count values, but also parameters indicating the relationship between the pixel of interest and the object signals of surrounding pixels. . Specifically, the count value Cnt1 is a parameter that includes the size of the object to which the target pixel belongs. The count values Cnt2 and Cnt3 are parameters that include the degree of isolation of the object to which the target pixel belongs.

  As described above, when the comprehensive determination unit 104 performs determination using the count values Cnt1, Cnt2, and Cnt3, it is possible to extract only objects that are isolated from surrounding objects and that are not more than a certain size. Become.

  That is, the object 521 shown in FIG. 5 conforms to all the equations 1 to 3, is isolated from other surrounding objects, and is equal to or smaller than a certain size, and thus is detected as an impurity.

  On the other hand, an object 522 that is a decimal point or a period of a part of a character is determined not to be an impurity from the standpoint of isolation because it does not conform to Equation 3 and is in an arrangement state close to other surrounding character objects.

  Further, since the pixel near the center of the object 523 that is a part of the alphabet “i” in the text conforms to Equation 3, there is a possibility that no character object exists within a certain range from the standpoint of isolation. . However, since the object 523 does not conform to Equation 1, it can be seen that the object 523 is larger than a certain size. Therefore, the object 523 can be determined to be a meaningful object such as a character or graphic from the viewpoint of size (that is, an object drawn by the creator of the input image with some intention) (that is, with impurities) Is determined to be not).

In addition, since the object 524 conforms to Equation 1 and Equation 2 , there is a possibility that the object 524 is an impurity from the viewpoint of size and isolation. However, the object 524 does not conform to Equation 3 , and in the second pixel group and the third pixel group around the object 524, small objects are in a state of being dispersed at a constant interval or randomly. I understand. For this reason, since the area including the object 524 has a structure composed of halftone dots such as photographs and graphics, the object 524 can be determined not to be an impurity from the standpoint of isolation.

  As described above, according to the processing described with reference to FIGS. 5 to 8, it is determined that the target pixel is not an impurity when the following three conditions are satisfied. That is, the count value Cnt1 (first count value) does not exceed the value set in advance as the size that the impurity can take. A value obtained by adding a count value Cnt2 (second count value) and a count value Cnt3 (third count value) to a value set in advance to determine that no object exists in the vicinity of the object to which the target pixel belongs If does not exceed. When the third count value does not exceed a value set in advance to determine that an object smaller than the neighboring object does not exist in the third pixel group.

  In the present embodiment, as described above, in the processing blocks 251 to 254, the target pixel and the surrounding pixels are divided into three pixel groups so as to form a layer around the target pixel, and the object signal is counted. I'm taking it. Hereinafter, the difference in effect between the case where the pixel of interest and the surrounding pixels are divided into three pixel groups as in the present embodiment and the case where the pixel is divided into two pixel groups will be described.

  FIG. 9 shows an example of an arrangement pattern in a case where the target pixel 902 and its peripheral pixels are divided into pixel groups 911 and 912 around the target pixel 902. Here, it is assumed that the count value of the pixel whose object signal is ON in the first inner pixel group 911 is Cnt1 '. The count value of the pixel whose object signal is ON in the second outer pixel group 912 is Cnt2 ′.

  In such a configuration, when an object signal 1001 with impurities as shown in FIG. 10 and an object signal 1002 with a part of a character are given, Cnt1 ′ and Cnt2 ′ are used to determine whether these objects are impurities. Consider how to use and determine. In the object signals 1001 and 1002, the count values Cnt1 'and Cnt2' for the target pixel 1011 are equal to the count values Cnt1 'and Cnt2' for the target pixel 1012 (that is, Cnt1 '= 36 Cnt2' = 15). As a result, the difference between the two objects cannot be identified using the count values Cnt1 'and Cnt2' no matter what conditional expression is set in step 241 for performing the comprehensive determination. That is, when the two-layer structure is adopted as in the configuration in this case, the impurity detection device may not be able to distinguish between impurities and those that are not. For this reason, a pattern having three or more layer structures is effective in realizing this embodiment.

  That is, in the present embodiment, the pixel of interest and its surrounding pixels are divided into three pixel groups, and Cnt2 'is divided into Cnt2 and Cnt3 and counted. As a result, when the arrangement pattern 401 of FIG. 4 is used, Cnt3 is 3 for the object corresponding to the signal 1001, and Cnt3 is 15 for the object corresponding to the signal 1002, so that both objects can be distinguished.

Here, the significance of counting Cnt1, Cnt2 and Cnt3 in this embodiment will be described in more detail. First, the impurities contained in the original paper are small. Therefore, in order for the target pixel to be a part of the impurity, the number of pixels (Cnt1) that are part of the object in the first pixel group is small. In addition, when the value of Cnt2 + Cnt3 is a certain large value, the target pixel may not be a part of the impurity. In other words, it is conceivable that the object including the target pixel is not an impurity because it has a large size, or the object including the target pixel is small but not an impurity (for example, a part of “i”, a period). Therefore, in order for the target pixel to be a part of the impurity, the value of Cnt2 + Cnt3 needs to be a small value to some extent. Further, even somewhat smaller values of Cnt1 and C nt 3, if small objects are not impurities such as scattered (e.g., dot newspapers) have. Therefore, in order for the target pixel to be a part of the impurity, Cnt2 + Cnt3 needs to be a value smaller than the number of pixels of such a small object. That is, when all of the above Equations 1 to 3 are satisfied, it is determined that the target pixel is a part of the impurity.

  On the other hand, as a modification, when using the pattern 1101 and the pattern 1102 having four or more pixel groups as shown in FIG. 11, although the processing load increases, a more accurate count value can be obtained. There is an effect of increasing the accuracy of the comprehensive determination. That is, in one or a plurality of pixel groups including pixels existing outside the third pixel group, the number of pixels that are part of the object is counted as the count value of each pixel group for each pixel group. Is done. Using all the counted values including the first to third count values, the size and the isolation degree of the object to which the target pixel belongs are determined. By determining whether or not the pixel of interest is an impurity in the input image based on the determined size and isolation of the object, the accuracy of impurity detection can be further improved.

  Further, as in the pattern 1103, at least a part of the four or more pixel groups is divided according to the direction from the target pixel, and each divided pixel group is a part of the object. You may count the number of By performing such counting, it is possible to give directivity to the comprehensive determination (detection of impurities) by the comprehensive determination unit 104. Alternatively, at least a part of the three pixel groups may be divided according to the direction from the target pixel, and the above-described counting and determination may be performed. In other words, the vertical direction has a low probability of the presence of adjacent character objects, so the overall determination threshold is increased to increase the sensitivity of impurity detection, and the horizontal direction has a high probability of the presence of adjacent character objects to reduce the sensitivity. It is possible to take a configuration such that Thereby, it is possible to prevent a character from being erroneously detected as an impurity while improving the capability of detecting impurities.

  As described above, in the present embodiment, it is possible to realize impurity detection using only input image data and performing low-load processing. In particular, when impurities such as those contained in a sheet are detected, it is possible to distinguish the impurities from small-point characters and symbols, thereby realizing highly accurate impurity detection.

  In the first embodiment, in the processing of the processing blocks 251 to 253, in order to calculate the count values Cnt1, Cnt2, and Cnt3, the object flags of all the reference pixels are repeatedly read for each pixel. However, in the present embodiment, a description will be given of a configuration in which the number of times of reading is reduced and an equivalent effect can be obtained with a lower load process.

  The configuration of the impurity detection apparatus in the present embodiment is similar to that of the first embodiment, and is as shown in FIG.

  FIG. 12 shows an overall operation flow by the impurity detection apparatus 100 in the present embodiment. Note that the processing order of the pixels in the image is assumed to be processed from left to right for each row as shown in FIG.

  FIG. 13 shows an arrangement pattern of reference pixels in the present embodiment. In the present embodiment, as in the first embodiment, reference pixels for the target pixel 1300 are divided into three pixel groups: a first pixel group 1301, a second pixel group 1302, and a third pixel group 1303. In addition, six pixels of a fourth pixel group 1304, a fifth pixel group 1305, a sixth pixel group 1306, a seventh pixel group 1307, an eighth pixel group 1308, and a ninth pixel group 1309 are added. Set the group.

  In the second embodiment, the difference from the first embodiment is that Cnt1, Cnt2, and Cnt3 are not calculated for each pixel, but a difference calculation is performed on each count value obtained in the pixel adjacent to the left of the target pixel. This is a point calculated as each count value of the target pixel.

Hereinafter, processing by the impurity detection apparatus 100 will be described with reference to FIG.
Since the processing in step 1201 and step 1202 is the same as that in step 201 and step 202, description thereof will be omitted.

  In processing block 1203 to processing block 1208, the characteristic acquisition unit 103 calculates the count value of the object signal of the pixel in the leftmost column in the image. That is, the count value Cnt1 of the first pixel group, the count value Cnt2 of the second pixel group, the count value Cnt3 of the third pixel group, the count value Cnt4 of the fourth pixel group, and the count value of the fifth pixel group Cnt5 and the count value Cnt6 of the sixth pixel group are counted. These counted values are stored in the storage unit 112 in association with the target pixel.

  The processing block 1251 includes steps 1209 to 1211. In this processing block, the comprehensive determination unit 104 performs comprehensive determination using the count value under the control of the main control unit 111.

  First, in step 1209, it is determined whether the object signal of the target pixel is ON. If the object signal is ON, the process transitions to step 1210. If the object signal is OFF, the process transitions to step 1212.

  Since the processing in step 1210 and step 1211 is the same as that in step 241 and step 242, description thereof will be omitted. The comprehensive determination unit 104 performs a comprehensive determination on the target pixel using the count values Cnt1, Cnt2, and Cnt3. If TRUE, the impurity flag is turned ON, and if FALSE, the target pixel is stored in association with the target pixel. The data is stored in the unit 112.

  In step 1212, the main control unit 111 calculates the count values shown in Equations 4 to 6 for the count values stored in the storage unit 112.

  In step 1213, the main control unit 111 resets the count values Cnt4, Cnt5, and Cnt6 stored in the storage unit 112 to zero. That is, in combination with step 1233 described later, the count values Cnt4, Cnt5, and Cnt6 are reset to zero each time a transition is made to a new target pixel loop. Specifically, each transition from step 1212 to step 1221, transition from step 1232 to step 1221, and transition from step 1232 to step 1237 is reset to zero.

  In step 1221, the target pixel is changed to the pixel on the right side, and impurities are subsequently determined for the new target pixel.

  In the processing block 1222 to the processing block 1227, the characteristic acquisition unit 103 calculates the count value of the object signal of the pixels other than the leftmost column in the image. That is, the count value Cnt4 of the fourth pixel group, the count value Cnt5 of the fifth pixel group, the count value Cnt6 of the sixth pixel group, the count value Cnt7 of the seventh pixel group, and the count value of the eighth pixel group Cnt8 and the count value Cnt9 of the ninth pixel group are calculated. The calculated count value is stored in the storage unit 112 in association with the target pixel.

  In step 1228, the main control unit 111 calculates the count values shown in Expression 7 to Expression 9 with respect to the count values stored in the storage unit 112.

  The processing block 1252 includes steps 1209 to 1211. In this processing block, the comprehensive determination unit 104 performs comprehensive determination using the count values Cnt1, Cnt2, and Cnt3 under the control of the main control unit 111. Since the processing is the same as the processing block 1251, the description thereof is omitted.

  In step 1232, the main control unit 111 calculates the count values shown in Equations 4 to 6 for the count values stored in the storage unit 112. That is, the same processing as in step 1212 is performed.

  In step 1233, the main control unit 111 resets the count values Cnt4, Cnt5, Cnt6, Cnt7, Cnt8, and Cnt9 stored in the storage unit 112 to zero. That is, in addition to the count values Cnt4, Cnt5, and Cnt6, Cnt7, Cnt8, and Cnt9 are also reset to 0 each time a transition is made to a new target pixel loop. That is, the value is reset to 0 each time the transition from step 1232 to step 1221 and the transition from step 1232 to step 1237 is performed.

  In step 1234, the main control unit 111 determines whether or not the target pixel is the right end of the image. If it is not the right end, the process proceeds to Step 1221. If it is the right end, the process transitions to step 1235.

  In step 1235, the main control unit 111 determines whether or not the target pixel is the lower end of the image. If it is not the lower end, the process proceeds to step 1236.

  In step 1236, the main control unit 111 resets the count values Cnt1, Cnt2, and Cnt3 stored in the storage unit 112 to zero. That is, the count values Cnt1, Cnt2, and Cnt3 are reset to zero each time a transition is made to a new line (each time the process transitions from step 1235 to step 1237).

  In step 1237, the target pixel is changed to the pixel at the left end one row below, and the impurity is determined for the new target pixel thereafter.

  If it is determined in step 1235 that the target pixel is at the lower end, the target pixel is at the lower right end, and the process ends.

The above is the configuration of this embodiment related to the impurity detection apparatus.
In this embodiment, as described above, Cnt1, Cnt2, and Cnt3 are not calculated for each pixel as in the first embodiment. Cnt4, Cnt5, Cnt6, Cnt7, Cnt8, and Cnt9 are calculated as differences from the respective count values obtained at the left adjacent pixel, and the count value of the target pixel is calculated by adding and subtracting these. Thereby, the frequency | count of reading the object signal memorize | stored in the memory | storage part 112 decreases. For example, when the arrangement pattern shown in FIG. 13 is used, in the embodiment described in the first embodiment, the object signal needs to be read 441 times for each pixel, but in this embodiment, only 94 times are required. In other words, it is possible to realize an impurity detection device having a precision equivalent to that of the first embodiment with a lower load process.

(Other examples)
The present invention can also be realized by executing the following processing. That is, software (program) that realizes the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, or the like) of the system or apparatus reads the program. It is a process to be executed. In this case, the program and the recording medium on which the program is recorded constitute the present invention.

DESCRIPTION OF SYMBOLS 100 Impurity detection apparatus 101 Image input part 102 Object detection part 103 Character acquisition part 104 General determination part 105 Main control part 106 Memory | storage part 107 Operation part 411 1st pixel group 412 2nd pixel group 413 3rd pixel group 421 Attention Pixel 521 Impurity 522 Decimal point 523 Part of character 524 Newspaper dot 902 Target pixel 911 First pixel group 912 Second pixel group 1011 Target pixel 1012 Target pixel 1300 Target pixel 1301 First pixel group 1302 Second pixel Group 1303 3rd pixel group 1304 4th pixel group 1305 5th pixel group 1306 6th pixel group 1307 7th pixel group 1308 8th pixel group 1309 9th pixel group

Claims (8)

Of the first pixel group consisting of the target pixel in the object in the input image and the peripheral pixels including the neighboring pixels of the target pixel , the number of pixels that are part of the object is counted as the first count value. Means to
Means for counting, as a second count value, the number of pixels that are part of the object among the second pixel group consisting of peripheral pixels including pixels close to the outside of the first pixel group;
Of the third group of pixels consisting of peripheral pixels including a pixel adjacent to the outer side of the second pixel group, and means for counting the number of pixels that are part of the object as a third count value,
Wherein based on the first count value and the second count value and the third count value, the pixel of interest and a means for determining whether the impurities in the input image,
The means for determining is
At least the first count value does not exceed the first threshold value, the third count value does not exceed the second threshold value, and at least the second count value and the third count value If the value based on the sum with bets does not exceed a third threshold value, an impurity detecting apparatus wherein the pixel of interest characterized that you determined to be impurities. In one or a plurality of pixel groups including pixels existing outside the third pixel group, the number of pixels that are part of the object is counted as a count value of each pixel group. With means,
Said determining means, on the basis of the counted every count value, impurity detecting apparatus according to claim 1, wherein the target pixel and judging whether the impurities in the input image. The determination means counts the number of pixels that are part of the object, which is counted for each pixel group obtained by dividing at least a part of the pixel group in the direction from the target pixel. The impurity detection device according to claim 2 , wherein the determination is further performed. The input image using the binarization result of the impurity detector according to any one of claims 1 to 3, characterized in that it comprises means for detecting the object in the input image. The input image using the result of differentiating the impurity detecting apparatus according to any one of claims 1 to 3, characterized in that it comprises means for detecting the object in the input image. Using the result of the image area separation for the input image, an impurity detecting apparatus according to any one of claims 1 to 3, characterized in that it comprises means for detecting the object in the input image . Of the first pixel group consisting of the target pixel in the object in the input image and the peripheral pixels including the neighboring pixels of the target pixel , the number of pixels that are part of the object is counted as the first count value. And steps to
Counting, as a second count value, the number of pixels that are part of the object among the second pixel group consisting of peripheral pixels including pixels close to the outside of the first pixel group;
Of the third group of pixels consisting of peripheral pixels including a pixel adjacent to the outer side of the second pixel group, a step of counting the number of pixels that are part of the object as a third count value,
Wherein based on the first count value and the second count value and the third count value, the pixel of interest and a step of determining whether the impurities in the input image,
In the determination,
At least the first count value does not exceed the first threshold value, the third count value does not exceed the second threshold value, and at least the second count value and the third count value If the value based on the sum with bets does not exceed a third threshold value, an impurity detecting method the pixel of interest characterized that you determined to be impurities. A program for causing a computer to execute the impurity detection method according to claim 7 .