JP2010134564A - Image processor, image processing system, image processing method, image processing program and recording medium recording the image processing program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image processor, an image processing system, an image processing method, an image processing program and a recording medium recording the image processing program, for determining whether or not a circumferential image is present with high accuracy even if storage capacity of the image processor is small. <P>SOLUTION: The image processor 101 is provided with: a template characteristic storage part 3 storing characteristics of a decision template; an estimation value calculation part 4 calculating a circle possibility area and an estimated circle center in an input image by use of the characteristics; an ideal gradient calculation part 5 calculating an ideal gradient from the estimated circle center and coordinates of interest in the circle possibility area; an approximate gradient calculation part 6 calculating an approximate gradient from the coordinates of interest in a specific image in the circle possibility area and coordinates that are coordinates in the specific image, close to the coordinates of interest; a gradient comparison part 7 outputting a gradient comparison result between the ideal gradient and the approximate gradient; and a circular shape decision part 8 performing decision based on the gradient comparison result. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

The present invention relates to an image processing apparatus, an image processing system, an image processing method, an image processing program, and a recording medium on which an image processing program is recorded, which makes it possible to determine whether or not a circular shape is present.

  Seals are widely used for roles such as proof of legitimacy of private documents and official documents, and identification of procedures. By the way, in order to determine the legitimacy of a document whose stamp has been pressed, it is necessary to determine whether or not the seal stamp used for creating the document is the same as that registered in advance. This determination is actually performed by comparing the imprint in the document with a pre-registered imprint and determining whether or not the two imprints match. Since it is very inefficient for humans to perform all of this collation work, various devices for assisting and supporting collation work have been developed.

  Patent Document 1 describes an apparatus for storing a partial pattern of a specific imprint and determining whether the pattern exists in an input image in order to prevent copying of an image including the specific imprint. .

JP 2000-76458 A

  However, in the apparatus described in Patent Document 1, since determination is performed by storing only a part of the imprint pattern, the determination accuracy is reduced as compared with the case where determination is performed by storing all imprints, and other than the specific imprint. Will be misjudged.

  In order to prevent such a misjudgment, it is first determined whether or not there is a circumferential portion in the specific imprint, and if it is determined that there is a circumferential portion thereafter, a precise imprint using the specific imprint is performed. A method for performing the determination is conceivable. With such a method, determination can be performed with high accuracy while speeding up the determination process. However, even in the determination of the circumference portion, there is a problem that the storage capacity becomes enormous if the image information of the entire circumference is stored, and the determination is made only with a part of the image information of the circumference. However, there is a problem that sufficient accuracy cannot be obtained.

  The present invention has been made to solve the above-described problem, and enables image processing to accurately determine whether or not a circumferential image exists even when the storage capacity of the image processing apparatus is small. An object is to provide an apparatus, an image processing system, an image processing method, an image processing program, and a recording medium on which the image processing program is recorded.

The present invention comprises a template information storage means for storing information of a template image;
Estimated circle center calculation means for calculating an estimated circle center, which is a coordinate estimated that the center of the circle exists in the input image, using the information;
A circle existence possibility area calculating means for calculating, using the information, a circle existence possibility area that is an area in which the circle may exist in the input image;
An ideal gradient calculating means for calculating an ideal gradient, which is a gradient of a tangent of a circle passing through the target coordinate, with the estimated circular center as the center of the circle from the estimated circle center and the target coordinate in the circle existence possibility region When,
An approximate tangent gradient at the target coordinate of the specific image is calculated from the target coordinate in the specific image in the circle existence possibility region and the coordinate in the specific image and in the vicinity of the target coordinate. An approximate gradient calculating means for calculating an approximate gradient;
A gradient comparison means for comparing the ideal gradient and the approximate gradient having the same target coordinates used for calculating the gradient and outputting a gradient comparison result;
An image processing apparatus comprising: a circular shape presence / absence determining unit that determines whether or not a circumferential image is included in the input image based on the gradient comparison result.

Further, the present invention comprises an estimated circumference calculating means for calculating an estimated circumference area that is a circumference area centered on the estimated circle center and that is an area within the circle existence possibility area,
The ideal gradient calculating means calculates an ideal gradient that is a gradient of a tangent of a circle passing through the target coordinate from the estimated circle center and the coordinate of interest in the estimated circumferential region, with the estimated circle center as the center of the circle. Calculate
The approximate gradient calculating means approximates the specific image at the target coordinate from the target coordinate in the specific image within the estimated circumferential area and the coordinate in the specific image and in the vicinity of the target coordinate. An approximate gradient that is a gradient of a typical tangent is calculated.

According to the present invention, the approximate gradient calculating unit includes a specific edge extracting unit that extracts a specific edge pixel that is a specific edge pixel in the circle existence possibility region, and a specific edge storage that stores coordinates of the specific edge pixel. Means, and first difference calculation means for calculating a coordinate difference between the specific edge pixels,
The specific image is an image including the specific edge pixels, and a target coordinate used for calculating an approximate gradient is selected from the specific edge pixels.

  Further, according to the present invention, the ideal gradient calculation means is a second difference calculation means for calculating a coordinate difference between the coordinates in the circle existence possibility area and the estimated circle center, and converts the coordinate difference to obtain the estimated circle center. And an ideal tangent vector calculation means for calculating a tangent vector of the circumference passing through the coordinates in the circle existence possibility region.

  Further, according to the present invention, the circular shape presence / absence determining means includes a gradient comparison result storage means for storing the gradient comparison result, and a circumferential image is included in the input image based on the gradient comparison result and a circumference determination threshold value. And a circumference determination means for determining whether or not the

  The present invention is further characterized by comprising a circumference determination threshold value changing means for changing the circumference determination threshold value.

The present invention also includes filter means for extracting an image of a specific color from the input image,
An image of a specific color extracted by the filter means is used as an input image, and it is configured to be able to determine whether or not a circumferential image is included in the input image.

The present invention also includes binarization processing means for performing binarization processing using information on the image of the specific color and information on the image other than the specific color,
An image after binarization processing by the binarization processing means is used as an input image, and it is configured to be able to determine whether or not a circumferential image is included in the input image. .

  According to another aspect of the present invention, there is provided an image processing system comprising: the image processing device; and a notification device that notifies a user of a result of determination by the image processing device using at least one of sound and light. .

The present invention also includes an estimated circle center calculation step of calculating an estimated circle center, which is a coordinate estimated to have a circle center in the input image, using information of the template image;
A circle existence possibility area calculating step for calculating a circle existence possibility area that is an area in which the circle may exist in the input image using the information;
An ideal gradient calculation step of calculating an ideal gradient that is a gradient of a tangent of a circle passing through the coordinate of interest, with the estimated circle center being the center of the circle from the estimated circle center and the coordinate of interest in the circle existence possibility region When,
An approximate tangent gradient at the target coordinate of the specific image is calculated from the target coordinate in the specific image in the circle existence possibility region and the coordinate in the specific image and in the vicinity of the target coordinate. An approximate gradient calculating step for calculating an approximate gradient;
A gradient comparison step of comparing the ideal gradient and the approximate gradient having the same target coordinates used for calculating the gradient and outputting a gradient comparison result;
And a circular shape presence / absence determining step for determining whether or not a circumferential image is included in the input image based on the gradient comparison result.

The present invention also provides a computer,
Template information storage means for storing information of the template image;
Estimated circle center calculation means for calculating an estimated circle center, which is a coordinate estimated that the center of the circle exists in the input image, using the information;
A circle existence possibility area calculating means for calculating, using the information, a circle existence possibility area that is an area in which the circle may exist in the input image;
An ideal gradient calculating means for calculating an ideal gradient, which is a gradient of a tangent of a circle passing through the target coordinate, with the estimated circular center as the center of the circle from the estimated circle center and the target coordinate in the circle existence possibility region When,
An approximate tangent gradient at the target coordinate of the specific image is calculated from the target coordinate in the specific image in the circle existence possibility region and the coordinate in the specific image and in the vicinity of the target coordinate. An approximate gradient calculating means for calculating an approximate gradient;
A gradient comparison means for comparing the ideal gradient and the approximate gradient having the same target coordinates used for calculating the gradient and outputting a gradient comparison result;
An image processing program that functions as circular shape presence / absence determining means for determining whether or not a circumferential image is included in the input image based on the gradient comparison result.

  The present invention also provides a computer-readable recording medium on which the program is recorded.

  According to the present invention, it is possible to calculate the estimated circle center and the circle existence possibility region using the information of the template image, and using the calculated estimated circle center and the circle existence possibility region, the ideal gradient and the approximate gradient are calculated. It is possible to calculate whether or not a circumferential image is included in the input image by comparing the calculated ideal gradient with the approximate gradient. The template image does not need to be an image including the entire circumference, and the estimated circle center and the circle existence possibility region can be calculated as long as the image includes a part of the circumference. Therefore, since the determination can be performed without storing the information on the entire circumferential image, the determination can be performed even if the template information storage unit is a small-capacity storage unit. Further, since the determination is performed by comparing the template image information and the ideal gradient with the approximate gradient, it is possible to perform the determination with higher accuracy than the apparatus that performs the determination using only the template image information.

  Further, according to the present invention, an estimated circumferential area can be calculated, an ideal gradient can be calculated from the estimated circle center and a target coordinate in the estimated circumferential area, and a specific image in the estimated circumferential area can be calculated. The approximate gradient can be calculated from the target coordinates and the coordinates in the specific image and in the vicinity of the target coordinates. Therefore, since the comparison between the ideal gradient and the approximate gradient is limited to the gradient with respect to the target coordinate in the estimated circumferential region, the processing of the image processing apparatus can be speeded up.

  Further, according to the present invention, it is possible to extract specific edge pixels in a circle existence possibility region and calculate an approximate gradient from the coordinate difference between the extracted specific edge pixels and the target coordinate.

  Further, according to the present invention, a coordinate difference is calculated from the estimated circle center and the coordinates in the circle existence possibility region, the estimated circle center is set as the center of the circle from the calculated coordinate difference, and the coordinates in the circle existence possibility region are calculated. The tangent vector of the circumference passing through can be calculated, and the ideal gradient can be calculated from the calculated tangent vector.

  Further, according to the present invention, it is possible to determine whether or not a circumferential image is included in the input image based on the gradient comparison result and the circumferential determination threshold value.

  Further, according to the present invention, the circumference determination threshold can be changed. Therefore, for example, even if the characteristics of the input image change due to aging degradation of the image reading device, the determination of the presence or absence of a circular shape can be made more than a certain accuracy by changing the circumference determination threshold according to the change in characteristics. Can keep.

  Further, according to the present invention, it is possible to extract an image of a specific color from an input image, and determine the presence or absence of a circular shape using the extracted image of the specific color as an input image. Therefore, when the circular color distribution of the determination target is known in advance, an image having a color different from that of the determination target can be cut by the filter unit, and the determination accuracy can be improved. Even if the input image is a color image, the determination can be made with high accuracy. Furthermore, since the new input image has less information, the determination can be made even if the image processing apparatus has a smaller configuration.

  Further, according to the present invention, the binarization processing means can perform binarization processing such that, for example, “1” corresponds to an image of a specific color and “0” corresponds to an image other than the specific color. The presence / absence of a circular shape can be determined using the image after the value processing as an input image. Therefore, since the new input image has little information, the determination can be made even if the image processing apparatus has a smaller configuration.

  Further, according to the present invention, the determination result by the image processing device can be notified to the user using at least one of sound and light by the notification device.

  According to the present invention, the estimated circle center and the circle existence possibility region are calculated using the template image information, and the ideal gradient and the approximate gradient are calculated using the calculated estimated circle center and the circle existence possibility region. It is possible to determine whether or not a circumferential image is included in the input image by comparing the calculated ideal gradient with the approximate gradient. The template image does not need to be an image including the entire circumference, and the estimated circle center and the circle existence possibility region can be calculated as long as the image includes a part of the circumference. Therefore, the determination can be performed without storing information on the entire circumferential image. Further, since the determination is performed by comparing the template image information and the ideal gradient with the approximate gradient, it is possible to perform the determination with higher accuracy than the apparatus that performs the determination using only the template image information.

  In addition, according to the present invention, the function of each unit described above can be realized by a computer, and the same effect as the above-described image processing apparatus can be achieved.

  Further, according to the present invention, the function of each unit described above can be realized in the computer by causing the computer to read it.

  An image processing apparatus 101 according to the first embodiment of the present invention will be described. The image processing apparatus 101 is an apparatus that determines whether or not a circumferential image exists in a document. For example, the image processing apparatus 101 can determine whether or not the circumferential image 32 in the specific imprint image 31 shown in FIG. In addition, the image processing apparatus 101 determines not only whether or not there is a perfect circle-shaped circumferential image 32 as shown in FIG. It is also possible to determine whether or not a shape circumferential image exists (hereinafter referred to as “ellipse shape presence / absence determination”). First, the image processing apparatus 101 will be described along a process for determining the presence / absence of a perfect circle shape (hereinafter referred to as a “perfect circle shape presence / absence determination process”).

  FIG. 1 is a block diagram functionally showing the configuration of the image processing apparatus 101. FIG. 2 is a block diagram functionally showing the configuration of the estimated value calculation unit 4 in the image processing apparatus 101. FIG. 3 is a block diagram functionally showing the configuration of the ideal gradient calculation unit 5 in the image processing apparatus 101. FIG. 4 is a block diagram functionally showing the configuration of the approximate gradient calculation unit 6 in the image processing apparatus 101. FIG. 5 is a block diagram functionally showing the configuration of the circular shape presence / absence determining unit 8 in the image processing apparatus 101. The image processing apparatus 101 includes a preprocessing unit 1, an imprint storage unit 2, a template feature storage unit 3 that is a template information storage unit, an estimated circle center calculation unit, and an estimated value calculation unit that is a circle existence possibility region calculation unit. 4, an ideal gradient calculation unit 5 as an ideal gradient calculation unit, an approximate gradient calculation unit 6 as an approximate gradient calculation unit, a gradient comparison unit 7 as a gradient comparison unit, and a circular shape presence / absence determination unit The determination unit 8 is provided and is electrically connected to the image reading device 9. The term “functional” does not necessarily mean that each member is independent, but storage means, calculation means, and control means that enable the image processing apparatus 101 to perform functions corresponding to each member. Is provided. Below, the function of each member is demonstrated in detail.

  The image reading device 9 is an image reading device such as a CCD (Charge Coupled Device), reads a reflected light image from a document image as an RGB (Red Green Blue) analog signal, and preprocesses the read analog RGB signal as an input image. Output to part 1.

  The preprocessing unit 1 performs A / D (analog / digital) conversion on the input image to convert the analog RGB signal into a digital RGB signal. Moreover, conventionally well-known various pre-processing can be performed. For example, noise generated by shading correction processing for removing various distortions generated in the illumination system, the imaging system, and the imaging system of the image reading device 9 and the mechanical and optical characteristics of the image reading device 9 with respect to the digital RGB signal. Noise removal processing for removing the image, background removal processing for separating each region of the digital RGB signal with a certain value or more as a background region, and input image based on the digital RGB signal, character area, area gradation image region , And a region separation process for separation into any of the photographic regions can be performed. The pre-processing unit 1 also functions as a storage unit for input images, and can store input images after various processing. For example, when the document read by the image reading device 9 is an A4 size document, an A4 size input image is stored.

  The imprint storage unit 2 includes storage means such as an HDD (hard disk drive), and stores image information of the specific imprint image 31 shown in FIG. The imprint storage unit 2 may store image information of a true circle determination template and features of a true circle determination template, which will be described later, or may be a storage device outside the image processing apparatus 101.

  The template feature storage unit 3 stores the feature of the perfect circle determination template. The feature of the perfect circle determination template may be calculated based on the image information of the true circle determination template read from the seal imprint storage unit 2 or the true circle stored in the seal imprint storage unit 2. The characteristics of the circle determination template may be stored as they are.

  FIG. 6 is a diagram for explaining the characteristics of the perfect circle determination template. In FIG. 6, a predetermined upper limit value and lower limit value are set for luminance (or lightness, etc.), pixels whose luminance (or lightness, etc.) is greater than or lower than the upper limit value are white, and luminance (or lightness, etc.) is upper limit. Pixels smaller than the value and larger than the lower limit value are drawn in black. As shown in FIGS. 6A and 6B, the perfect circle determination template 33 is included in a part 34 of the specific imprint image 31 and includes a part of a perfect circle-shaped circumferential image 32. This is an image of the area. As shown in FIG. 6B, the perfect circle determination template 33 is divided into small areas (hereinafter referred to as “template small areas”).

  The scale in FIG. 6B is a scale indicating template small area coordinates representing the position of one small area in the perfect circle determination template 33, and the scale in FIG. 6C is a perfect circle determination template. 33 is a scale showing template pixel coordinates representing the position of one pixel within 33. FIG. 6C shows one template small region 35 (template small region whose template small region coordinates are (v, f)) composed of an image of a region of 20 pixels × 10 pixels. Further, the template small area may be composed of images of other sizes other than 20 pixels × 10 pixels.

  In the small template region, the number of pixels (hereinafter referred to as “predetermined pixels”) whose luminance (or brightness, etc.) is smaller than a predetermined upper limit value and larger than a predetermined lower limit value (pixels drawn in black in FIG. 6C) Is referred to as the feature amount of the template small region. As shown in FIG. 6C, the feature amount of the template small area 35 is 100.

  In addition, when all the template small areas are rectangles having the same shape as shown in FIG. 6B, the template pixel coordinates (hereinafter referred to as “template small area start point”) serving as the start point of the template small area are used. Each template small region coordinate can be represented as a representative. Further, a combination of the template small region start point and the feature amount can be calculated for all template small regions from one perfect circle determination template 33. A group consisting of such a combination of the template small region start point and the feature amount is referred to as a feature of the perfect circle determination template 33. As will be described later, since the feature of the perfect circle determination template 33 is used to calculate the estimated circle center and the circle existence possibility region in the input image, the perfect circle determination template 33 has a perfect circle circumference. The image need not be an image including all the images 32, and may be a minimum image that can calculate the estimated circle center and the circle existence possibility region. Considering the accuracy of calculation of the estimated circle center and the circle existence possibility region, a perfect circle determination template composed of an image including about a quarter of the circumferential image is sufficient.

  Further, the template feature storage unit 3 calculates the template pixel coordinates of one pixel in the true circle determination template 33 and the true circle from the one pixel for calculating the estimated circle center and circle existence possibility region described later. The main scanning direction difference and the sub-scanning direction difference to the center, and the main scanning direction range and the sub-scanning direction range indicating the minimum area including the perfect circle are stored.

  The estimated value calculation unit 4 includes a feature amount calculation unit 10, a template comparison history storage unit 11, a feature amount comparison unit 12, a feature amount comparison result storage unit 13, a match determination unit 14, an estimated circle center calculation unit 15, and a circle can exist. The sex region calculation unit 16 is included.

  The feature quantity calculation unit 10 determines a comparison area from input images including a perfect circle shape and an image similar to the perfect circle shape shown in FIG. 7A (a), and the perfect circle read from the template feature storage unit 3 The feature of the comparison area is calculated using the feature of the determination template 33. 7A, 7B, and 7C are diagrams for explaining the calculation of the feature amount by the feature amount calculation unit 10. FIG. 7A, 7B, and 7C, like FIG. 6, predetermined pixels are drawn in black, and pixels other than the predetermined pixels are drawn in white. The comparison areas 36 and 37 are images included in some of the images 39 and 40 of the input image 38, and are divided into small areas having the same shape as the perfect circle determination template 33, and the small areas are arranged in the same manner. Further, it is an image of the hatched portion. A small area in the comparison areas 36 and 37 is called a comparison small area.

  7B (b1) and 7C (b2) show the peripheral images 41 and 42 of the comparison regions 36 and 37, respectively. 7B (d1) and 7C (d2) show one comparative small area 43 and 44 in the comparison areas 36 and 37 shown in FIGS. 7B (c1) and 7C (c2), respectively. The scales in FIGS. 7B (c1) and 7C (c2) are scales indicating the comparative small area coordinates representing the position of one small area in the comparative areas 36 and 37, respectively. FIGS. 7B (d1) and 7C (d2) ) Is a scale indicating local pixel coordinates representing the position of one pixel in the comparison regions 36 and 37, respectively. The coordinates representing the position of one pixel in the input image 38 with respect to the local pixel coordinates are referred to as global pixel coordinates.

  In addition, when all the template small areas are rectangles having the same shape as shown in FIG. 6B, all the comparison small areas corresponding to the template small areas are also rectangles having the same shape. The local pixel coordinates (hereinafter referred to as “comparison small area start point”) can be representatively represented by the comparative small area coordinates. Therefore, the comparison area is an area where all the comparison small area coordinates coincide with the template small area coordinates, and all the comparison small area start points coincide with the template small area start point.

  The predetermined number of pixels in the comparative small area (the number of pixels drawn in black in FIGS. 7B (d1) and 7C (d2)) is referred to as the feature amount of the comparative small area. From FIG. 7B (d1) and 7C (d2), the characteristic amounts of the comparative small areas 43 and 44 are 100 and 96, respectively. In addition, the combination of the comparison small area start point and the feature amount can be calculated for all the comparison small areas from the comparison area. A group consisting of such a combination of the comparison small region start point and the feature amount is referred to as a comparison region feature. The feature amount calculation unit 10 outputs the calculated feature of the comparison area to the feature amount comparison unit 12.

  Further, the feature amount calculation unit 10 outputs a combination of local pixel coordinates and global pixel coordinates of one pixel in the comparison region where the feature is calculated to the template comparison history storage unit 11. If the combination of the local pixel coordinates and the global pixel coordinates is different, a different area in the input image 38 is used as the comparison area. As will be described later, when the determination that the perfect circle determination template does not match the comparison region is output from the matching determination unit 14, the feature amount calculation unit 10 calculates the features of different comparison regions. The feature amount calculation unit 10 is set so as not to perform the second feature calculation for the same comparison region from the combination of the local pixel coordinates and the global pixel coordinates read from the template comparison history storage unit 11.

  As described above, the template comparison history storage unit 11 stores a combination of local pixel coordinates and global pixel coordinates output from the feature amount calculation unit 10.

  The feature amount comparison unit 12 uses a combination of the template small region start point and the feature amount read from the template feature storage unit 3 and a combination of the comparison small region start point and the feature amount output from the feature amount calculation unit 10. Thus, it is determined whether or not the template small area and the comparative small area match. This determination is performed by comparing the feature amount of the template small region where the template small region start point and the comparison small region start point coincide with the feature amount of the comparative small region. For example, the template small region start point of the template small region 35 in FIG. 6C is (421, 51), and the comparative small region 43 in FIG. 7B (d1) and the comparative small region 44 in FIG. 7C (d2) are comparatively small. Since the region start point is (421, 51), the feature amount of the template small region 35 and the feature amount of the comparison small regions 44, 45 are compared.

  The feature amount comparison unit 12 determines that the small regions match when the feature amount of the comparison small region is within ± 10% of the feature amount of the template small region, and the feature amount ± 10 of the template small region. If not within%, it is determined that the small areas do not match. For example, the feature amount of the template small region 35 is 100, and the feature amount of the comparative small region in the case of the comparative small region 43 is 100. Therefore, the feature amount of the comparative small region is ± 0% of the feature amount of the template small region. Thus, it is determined that the template small area 35 and the comparative small area 43 match. Further, since the feature amount of the comparison small region in the case of the comparison small region 44 is 96, the feature amount of the comparison small region becomes -4% of the feature amount of the template small region. Is determined to match. These determination results are output to the feature amount comparison result storage unit 13.

  The feature quantity comparison result storage unit 13 is determined not to match the number determined to match (hereinafter referred to as “matching determination number”) from the determination result output from the feature quantity comparison unit 12. (Hereinafter referred to as “the number of mismatch determinations”). The sum of the coincidence determination number and the mismatch determination number is the number of small template regions (comparison small regions), and is the total number of determinations.

  The coincidence determination unit 14 reads out the determination result stored in the feature amount comparison result storage unit 13 and determines whether or not the perfect circle determination template matches the comparison region. This determination is performed using the coincidence determination number and the mismatch determination number. For example, if the number of coincidence determinations / the total number of determinations × 100 (%), which is the ratio of the number of coincidence determinations to the total number of determinations, is 10% or more, the match determination unit 14 matches the perfect circle determination template with the comparison region. Judge that you are doing. The determination result is output to the feature amount calculation unit 10, the estimated circle center calculation unit 15, and the circle existence possibility region calculation unit 16.

  When the determination that the true circle determination template matches the comparison area is output from the match determination unit 14, the estimated circle center calculation unit 15 calculates the estimated circle center. The estimated circle center is a global pixel coordinate that is estimated to have the center of the circle if the image in the comparison region is a part of the circumferential image. The estimated circle center calculation unit 15 reads the template pixel coordinates of one pixel in the perfect circle determination template read from the template feature storage unit 3 and the main scanning direction from the one pixel to the center of the perfect circle. The estimated circle center is calculated using the difference and the sub-scanning direction difference. FIG. 8 is a diagram for explaining estimated circle center calculation and circle existence possibility region calculation. As shown in FIGS. 8A1 and 8A2, the estimated circle center calculation unit 15 adds the main scanning direction difference 47, to the global pixel coordinates of the pixels 45, 46 having local pixel coordinates that coincide with the template pixel coordinates. 48 and the sub-scanning direction differences 49 and 50 are added to calculate estimated circle centers 51 and 52, respectively. The calculated estimated circle center is output to the difference vector calculation unit 17 of the ideal gradient calculation unit 5.

  When the determination that the true circle determination template matches the comparison region is output from the match determination unit 14, the circle presence possibility region calculation unit 16 calculates a circle presence possibility region. The circle existence possibility area is an area that is estimated to be a minimum area including the circumferential image if the image in the comparison area is a part of the circumferential image. The circle presence possibility area calculation unit 16 reads the template pixel coordinates of one pixel in the perfect circle determination template read from the template feature storage unit 3 and the main scanning direction indicating the minimum area including the perfect circle A circle existence possibility region is calculated using the range and the sub-scanning direction range. As shown in FIGS. 8B1 and 8B2, the circle presence possibility region calculation unit 16 calculates the global pixel coordinates of the pixels 45 and 46 having local pixel coordinates that match the template pixel coordinates, and the main scanning direction range. 53, 54 and the sub-scanning direction ranges 55, 56 are used to calculate circle existence possibility regions 57, 58. The calculated circle existence possibility region is output to the difference vector calculation unit 17 of the ideal gradient calculation unit 5 and the specific edge extraction unit 20 of the approximate gradient calculation unit 6.

  The ideal gradient calculation unit 5 includes a difference vector calculation unit 17 that is a second difference calculation unit, an ideal tangent vector calculation unit 18 that is an ideal tangent vector calculation unit, and a gradient calculation unit 19.

  The difference vector calculation unit 17 uses the estimated circle center output from the estimated circle center calculation unit 15 and the circle existence possibility region output from the circle existence possibility region calculation unit 16 to determine the global coordinates (hereinafter, “ And a difference vector between the center of the estimated circle and the center of the estimated circle. The target pixel in this case is one pixel in the circle existence possibility region, and is selected by the difference vector calculation unit 17. For example, when the estimated circle center is (Ox, Oy) and the target coordinate is (X, Y), the difference vector calculation unit 17 calculates (X-Ox, Y-Oy) as a difference vector. The calculated difference vector is output to the ideal tangent vector calculation unit 18 together with the target coordinate.

  The ideal tangent vector calculation unit 18 calculates an ideal tangent vector from the difference vector output from the difference vector calculation unit 17. The ideal tangent vector is a tangent vector when it is assumed that the target pixel is on the circumference centered on the estimated circle center. The ideal tangent vector is calculated as (ΔX, ΔY) = (Y−Oy, −X + Ox) by orthogonally transforming the difference vector. The calculated ideal tangent vector is output to the gradient calculation unit 19 together with the target coordinate.

  The gradient calculation unit 19 calculates an ideal gradient ΔY / ΔX = (− X + Ox) / (Y−Oy), which is the gradient of the ideal tangent vector, from the ideal tangent vector output from the ideal tangent vector calculation unit 18. The calculated ideal gradient is output to the gradient comparison unit 7 together with the target coordinate. When ΔX = Y−Oy = 0, the ideal gradient is not calculated, so a symbol such as “∞” is output instead of a number.

  The approximate gradient calculation unit 6 includes a specific edge extraction unit 20 that is a specific edge extraction unit, a specific edge storage unit 21 that is a specific edge storage unit, a difference vector calculation unit 22 that is a first difference calculation unit, and a gradient calculation unit 23. Including.

  The specific edge extraction unit 20 extracts all the global pixel coordinates of the specific edge pixels in the circle existence possibility region based on the circle existence possibility region output from the circle existence possibility region calculation unit 16. The specific edge pixel is an edge pixel that is not included in a closed region formed by other edge pixels among edge pixels extracted by a conventionally known method. FIG. 9 is a diagram for explaining extraction of specific edge pixels by the specific edge extraction unit 20. The specific edge extraction unit 20 extracts edge pixels from the images in the circle existence possibility regions shown in FIGS. 9A1 and 9A2 by a conventionally known method, and FIGS. A plurality of edge pixels 59 to 65 shown in b2) are extracted. The specific edge extraction unit 20 further extracts specific edge pixels 59 and 60 that are not included in the closed region formed by the other edge pixels from the plurality of edge pixels 59 to 65. The global pixel coordinates of the extracted specific edge pixel are output to the specific edge storage unit 21.

  The specific edge storage unit 21 stores all the global pixel coordinates of the specific edge pixels in the circle existence possibility area extracted as described above.

  The difference vector calculation unit 22 reads the global pixel coordinates of the specific edge pixel stored in the specific edge storage unit 21 and is a specific edge pixel having a main scanning direction difference of 3 or more between the target coordinate and the target pixel. A difference vector from the global pixel coordinate of the pixel having the closest distance is calculated. The target pixel in this case is one of the specific edge pixels, and is selected by the difference vector calculation unit 22. FIG. 10 is a diagram for explaining calculation of the difference vector by the difference vector calculation unit 22. For example, when the coordinate (X, Y) = (2005, 2011) of the pixel of interest 61 is 3 as shown in FIG. The global pixel coordinates (A, B) = (2008, 2007) of the pixel 62 that is the specific edge pixel and is closest to the target pixel are read out, and the difference vector 63 is (ΔX ′, ΔY ′) = (( (A−X), (B−Y)) = (3, −4) is calculated. Further, as shown in FIG. 10B, when the coordinate of the pixel of interest 64 is (X, Y) = (1008, 1009), it is a specific edge pixel whose main scanning direction difference from the pixel of interest 64 is 3 or more. Then, the global pixel coordinates (A, B) = (1008, 1005) of the pixel 65 closest to the target pixel are read out, and the difference vector 66 is (ΔX ′, ΔY ′) = ((A−X), ( B−Y)) = (0, −4) is calculated. The calculated difference vector is output to the gradient calculation unit 23 together with the target coordinate.

  The gradient calculation unit 23 calculates an approximate gradient ΔY ′ / ΔX ′ = (B−Y) / (A−X), which is the gradient of the difference vector, from the difference vector output from the difference vector calculation unit 22. The calculated approximate gradient is output to the gradient comparison unit 7 together with the target coordinate. When ΔX ′ = A−X = 0, the approximate gradient is not calculated, so a symbol such as “∞” is output instead of a number.

  The gradient comparison unit 7 uses the attention coordinate and ideal gradient output from the gradient calculation unit 19 and the attention coordinate and approximate gradient output from the gradient calculation unit 23 to determine whether the ideal gradient and the approximate gradient match. Judgment is made. This determination is performed by comparing the ideal gradient and the approximate gradient with the same coordinates of interest. FIG. 11 is a diagram schematically illustrating a comparison between the ideal gradient and the approximate gradient by the gradient comparison unit 7. As shown in FIG. 11A, the gradient comparison unit 7 includes an ideal gradient 71 that is a gradient of a tangent on a virtual circumference 69 passing through the target coordinate 67, and a specific edge pixel that is approximately calculated. The approximate gradient 73 that is the gradient of the tangent line is compared. Also, as shown in FIG. 11B, the gradient comparison unit 7 includes an ideal gradient 72 that is a gradient of a tangent on a virtual circumference 70 that passes through the target coordinate 68, and a specific edge pixel that is approximately calculated. Compare the approximate gradient 74, which is the gradient of the tangent line above. For example, when the approximate gradient is within the ideal gradient ± 10%, the gradient comparing unit 7 determines that the gradients match, and when the approximate gradient is not within ± 10%, the gradients do not match. Is determined. The determination result is output to the gradient comparison result storage unit 24 of the circular shape presence / absence determination unit 8.

  The circular shape presence / absence determination unit 8 includes a gradient comparison result storage unit 24 that is a gradient comparison result storage unit and a circumference determination unit 25 that is a circumference determination unit.

  The gradient comparison result storage unit 24 determines from the determination result output from the gradient comparison unit 7 that it does not match the number determined to match (hereinafter referred to as “gradient match determination number”). The number (hereinafter referred to as “gradient mismatch determination number”) is stored.

  The circumference determination unit 25 reads the determination result stored in the gradient comparison result storage unit 24, and determines whether or not a circumference image exists in the circle existence possibility region. This determination is performed using the gradient match determination number and the gradient mismatch determination number. The circumference determination unit 25 is, for example, the ratio of the gradient match determination number / the total gradient determination number × 100 (%), which is the ratio of the gradient match determination number to the gradient total determination number that is the sum of the gradient match determination number and the gradient mismatch determination number. Is equal to or greater than the circumference determination threshold value such as 10%, it is determined that the circumference image exists in the circle existence possibility area, and the circumference existence determination, the estimated circle center, and the circle existence possibility area are determined. The data is output to another member in the image processing apparatus 101 or another apparatus outside the image processing apparatus 101. In addition, the circle determination unit 25 determines that a circle existence possibility region where the number of gradient matching determinations / the total number of gradient determinations × 100 (%) is equal to or greater than the circle determination threshold does not exist in the input image. It is determined that the circumferential image does not exist, and the circumferential absence determination is output to another member in the image processing apparatus 101 or another apparatus outside the image processing apparatus 101.

  The true circle shape presence / absence determination process of the image processing apparatus 101 including the above-described members will be described. FIG. 12 is a flowchart showing the perfect circle shape determination process.

  The image processing apparatus 101 causes the image reading apparatus 9 to read a document image by a control unit (not shown) (step S1). The input image obtained by reading the document is preprocessed by the preprocessing unit 1. The image processing apparatus 101 determines whether or not there is an area that is not yet a comparison area in the input image after the preprocessing (step S2), and if there is, the process proceeds to step S3, and if not, the process proceeds to step S11. And proceed. When the comparison region is determined by the feature amount calculation unit 10 and the feature of the comparison region is calculated, the feature amount comparison unit 12 compares the feature of the perfect circle determination template with the feature of the comparison region (step S3). . Based on the result of this comparison, the match determination unit 14 determines whether or not the perfect circle determination template matches the comparison region (step S4). If they match, the process proceeds to step S5, and if they do not match, the process proceeds to step S2.

  For example, the characteristics of the comparison area 36 in FIG. 7B (c1) and the characteristics of the comparison area 37 in FIG. 7C (c2) are similar to the characteristics of the perfect circle determination template 33 in FIG. The processing apparatus determines that the perfect circle determination template 33 matches the comparison areas 36 and 37, and advances the process to step S5.

  The image processing apparatus 101 calculates the estimated circle center by the estimated circle center calculation unit 15 (step S5), and calculates the circle existence possibility region by the circle existence possibility region calculation unit 16 (step S6). The image processing apparatus 101 calculates an ideal gradient by the ideal gradient calculation unit 5 based on the calculated estimated circle center and circle existence possibility region (step S7). The image processing apparatus 101 calculates an approximate gradient by the approximate gradient calculation unit 6 based on the calculated circle existence possibility region (step S8).

  When the ideal gradient and the approximate gradient are calculated, the gradient comparison unit 7 compares the calculated ideal gradient with the approximate gradient (step S9). Based on this, the circular shape presence / absence determination unit 8 determines whether or not a circumferential image exists in the calculated circle existence possibility region (step S10).

  For example, as shown in FIG. 11 (a), when the ideal gradient 71 and the approximate gradient 73 match, and the ideal gradient and the approximate gradient similarly match in all the specific edge pixels, the calculation is performed. It is determined that a circumferential image exists in the circle existence possibility region. Further, as shown in FIG. 11B, when the ideal gradient 72 and the approximate gradient 74 do not coincide with each other, and the ideal gradient and the approximate gradient do not coincide with each other in many other specific edge pixels as well. Is determined that no circumferential image exists in the calculated circle existence possibility region. The result of determining whether or not a circumferential image exists in the calculated circle existence possibility area, and if there is a circumference image, the estimated circle center and the circle existence possibility area are stored in a storage unit (not shown). The

  The image processing apparatus 101 determines whether or not there is an area that is not yet a comparison area in the input image (step S11), and if it exists, the process proceeds to step S3, and if not, the process proceeds to step S12. . The image processing apparatus 101 uses the determination result stored in the storage means (not shown) to determine whether or not a circumferential image exists in the circle existence possibility area. If there is a circumferential image, the image processing apparatus 101 proceeds to step S13. If not, the process proceeds to step S14 (step S12).

  In step S <b> 13, the image processing apparatus 101 outputs a circumference existence determination, an estimated circle center, and a circle existence possibility area, which are information that a circumference image exists in the input image. If there are a plurality of circumference images in the input image, a plurality of circumference existence determinations are output. In step S <b> 14, the image processing apparatus 101 outputs a circumference nonexistence determination which is information that the circumference image does not exist in the input image.

  When the image processing apparatus 101 outputs the circumference presence determination or the circumference non-existence determination, the image processing apparatus 101 ends the perfect circle shape presence / absence determination process.

  The effects produced by the image processing apparatus 101 will be described. As described above, the image processing apparatus 101 determines the presence or absence of a perfect circle shape by using the comparison of the perfect circle determination template and the comparison region and the comparison of the ideal gradient and the approximate gradient. If the perfect circle determination template is an image including all of the circumferential images, it is possible to determine the presence or absence of a perfect circle shape with a certain degree of accuracy without comparing the ideal gradient and the approximate gradient. However, when the perfect circle determination template is an image including all of the circumference images, the amount of feature information of the perfect circle determination template becomes enormous, and therefore the template feature storage unit 3 needs to be a large-capacity storage unit. There is. That is, the image processing apparatus 101 according to the present invention can determine the presence or absence of a perfect circle shape even if the storage capacity of the template feature storage unit 3 is small.

  In addition, the image processing apparatus 101 performs the determination of the presence or absence of a perfect circle shape by using the comparison of the perfect circle determination template and the comparison region and the comparison of the ideal gradient and the approximate gradient, so the comparison between the perfect circle determination template and the comparison region is performed. The determination accuracy is improved as compared with the image processing apparatus that performs the determination only by the above. FIG. 13 is a diagram for explaining the effect of improving the determination accuracy in the image processing apparatus 101. FIG. 13A1 shows an image 75 in the comparison area where the circumference existence determination is output by the image processing apparatus 101, and FIG. 13A2 shows the circumference nonexistence determination output by the image processing apparatus 101. An image 76 in the comparison area is shown. 13B1 shows a comparative small area 77 in the image 75, and FIG. 13B2 shows a comparative small area 78 in the image 76.

  As shown in FIGS. 13 (a1), (a2), (b1), and (b2), there is a lack in the circumferential image in the comparative small regions 77 and 78 between the image 75 and the image 76. There is no difference in whether or not. Further, as shown in FIGS. 13B1 and 13B2, since the comparison small areas 77 and 78 have the same predetermined pixels, that is, feature amounts are equal, the comparison area including the image 75 and the comparison including the image 76 are compared. The region characteristics are the same. Therefore, in the image processing apparatus that performs the determination only by comparing the perfect circle determination template and the comparison area, not only the image 75 that is the circumferential image but also the image 76 in which the circumferential image is missing is the circumferential image. Determined.

  The image processing apparatus 101 of the present invention does not make an erroneous determination that the image 76 is a circumferential image. After comparing the perfect circle determination template and the comparison region, the image processing apparatus 101 calculates a circle existence possibility region and extracts specific edge pixels. FIGS. 13C1 and 13C2 are diagrams illustrating specific edge pixels in the comparative small areas 77 and 78. FIG. As shown in FIGS. 13C1 and 13C2, the specific edge pixel of the image 75 that is the circumferential image is only the pixel outside the circumferential image, whereas the image 76 in which the circumferential image is missing. The specific edge pixels are pixels on the outer side and the inner side of the circumferential image. Therefore, since the number of gradient matching determinations is reduced in the image 76, the image processing apparatus 101 does not determine that the image 76 is a circumferential image. Thus, the image processing apparatus 101 of the present invention can accurately determine the presence or absence of a perfect circle shape.

  Next, a process for determining the presence / absence of an elliptical shape in the image processing apparatus 101 (hereinafter referred to as an “elliptical shape determination process”) will be described. Similar to the perfect circle shape determination process, the input image obtained by reading the document is preprocessed by the preprocessing unit 1. In the image processing apparatus 101, the estimated value calculation unit 4 calculates the estimated circle center and the circle existence possibility region in the comparison region in the input image after the preprocessing. The determination template used in the calculation is an ellipse determination template. FIG. 14 is a diagram showing an ellipse determination template. In FIG. 14, a predetermined upper limit value and lower limit value are provided for luminance (or brightness, etc.), and luminance (or brightness, etc.) is greater than the upper limit value or lower than the lower limit value. Small pixels are drawn white, and pixels whose luminance (or brightness, etc.) is smaller than the upper limit value and larger than the lower limit value are drawn black. As shown in FIGS. 14 (a1), (b1), (a2), and (b2), the ellipse determination templates 83 and 84 are included in the parts 81 and 82 of the specific imprint images 79 and 80, and have an elliptical shape. This is an image of the shaded area including a part of the circumferential images 85 and 86. Further, as shown in FIGS. 14B1 and 14B2, the ellipse determination templates 83 and 84 are divided into template small regions. Similar to the perfect circle shape determination process, the features of the ellipse determination templates 83 and 84, which are groups of combinations of the feature amount of the template small region and the template small region start point, are stored in the template feature storage unit 3. Yes.

  The estimated value calculation unit 4 reads the feature of the ellipse determination template, compares the feature of the ellipse determination template with the feature of the comparison region, and determines whether or not they match. The estimated value calculation unit 4 calculates the estimated circle center and the circle existence possibility region based on the determination result. The image processing apparatus 101 calculates an ideal gradient by the ideal gradient calculation unit 5 and an approximate gradient by the approximate gradient calculation unit 6 based on the calculated estimated circle center and circle existence possibility region.

The method for calculating the ideal tangent vector by the ideal tangent vector calculation unit 18 in calculating the ideal gradient differs between the perfect circle shape presence / absence determination process and the elliptical shape presence / absence determination process. That is, assuming that the estimated circle center is (Ox, Oy) and the target coordinate is (X, Y), the difference vector calculated by the difference vector calculation unit 17 is (X−Ox, Y−Oy) = (A ′, B ′). ), And a formula representing a predetermined ellipse is x ² / C ² + y ² / D ² = 1 ((x, y) is an arbitrary point on the circumference of the ellipse, and C and D are constants) The ideal tangent vector (ΔX, ΔY) is calculated as (ΔX, ΔY) = (B ′ / D ² , −A ′ / C ² ). As shown in FIG. 14A2, when the inclination angle of the ellipse is α, the ideal tangent vector (ΔX, ΔY) is calculated as the following equation. The ideal gradient calculation unit 5 calculates an ideal gradient using such an ideal tangent vector. The calculation of the approximate gradient by the approximate gradient calculation unit 6 is the same as the perfect circle shape presence / absence determination process.

  When the ideal gradient and the approximate gradient are calculated, the gradient comparison unit 7 compares the calculated ideal gradient with the approximate gradient. The shape presence / absence determination unit 8 determines whether or not an elliptical circumferential image exists, and the determination result is output in the same manner as the perfect circle shape presence / absence determination processing. The above is the ellipse shape presence / absence determination process. The image processing apparatus 101 determines the circumference image portion in the imprint image even if an arbitrary imprint image exists at an arbitrary angle in the input image by the true circle shape presence / absence determination processing and the elliptical shape presence / absence determination processing. It can be performed.

  Next, an image processing apparatus 102 according to the second embodiment of the present invention will be described. FIG. 15 is a block diagram functionally showing the configuration of the image processing apparatus 102. As shown in FIG. 15, the image processing apparatus 102 has the same configuration as the image processing apparatus 101 except that the image processing apparatus 102 includes an estimated circumference calculation unit 26 that is an estimated circumference calculation unit. Therefore, only the functions of the estimated circumference calculation unit 26 and its peripheral members will be described. Each member common to the members of the image processing apparatus 101 has a common function.

  The estimated circumference calculation unit 26 stores a lower limit value of the circumference radius and an upper limit value of the circumference radius, the estimated circle center output from the estimated circle center calculation unit 15, and a circle existence possibility region. An estimated circumference area is calculated from the circle existence possibility area output from the calculation unit 16. FIG. 16 is a diagram for explaining the estimated circumferential region. As shown in FIG. 16A, the estimated circumference calculation unit 26 has a circumference having an estimated circle center 87 as a center and a circumference radius upper limit 88 as a radius, and an estimated circle center 87 as a center. An estimated circumference area 91 indicated by a hatched area, which is an area between the circumference having the value 89 as a radius and included in the circle existence possibility area 90, is calculated. The calculated estimated circumferential area is output to the difference vector calculation unit 17 and the specific edge extraction unit 20.

  The difference vector calculation unit 17 calculates a difference vector between the target coordinate and the estimated circle center from the estimated circle center output from the estimated circle center calculation unit 15 and the estimated circumference region output from the estimated circumference calculation unit 26. To do. The pixel of interest in this case is one pixel in the estimated circumferential area, and is selected by the difference vector calculation unit 17. The calculated difference vector is output to the ideal tangent vector calculation unit 18 together with the target coordinate. After the calculation of the difference vector, the ideal gradient is calculated and output to the gradient comparison unit 7 as in the image processing apparatus 101.

  The specific edge extraction unit 20 extracts all global pixel coordinates of the specific edge pixel in the estimated circumference region based on the estimated circumference region output from the estimated circumference calculation unit 26. FIGS. 16B and 16C show specific edge pixels extracted from the estimated circumferential area by the specific edge extraction unit 20. As shown in FIG. 16B, when all the specific edge pixels are present in the estimated circumferential area, all the specific edge pixels 92 in the circle existence possibility area are extracted. In addition, as shown in FIG. 16C, when there is a specific edge pixel 93 in the estimated circumferential area and a specific edge pixel 94 that is in the circle existence possibility area but outside the estimated circumferential area. In FIG. 16C, only the specific edge pixel 93 indicated by the solid line is extracted. After the specific edge pixel is extracted, an approximate gradient is calculated and output to the gradient comparison unit 7 as in the image processing apparatus 101. Thereafter, the determination is performed in the same manner as the image processing apparatus 101 from the output ideal gradient and approximate gradient.

  As described above, the image processing apparatus 102 can compare the gradients with respect to the target pixel in the estimated circumferential region, so that the determination processing speed can be increased while maintaining the determination accuracy. The image processing apparatus 102 also uses the parameters for designating the estimated circumferential area of the ellipse, for example, the upper limit value and the lower limit value of the major axis, the upper limit value and the lower limit value of the minor axis, and the inclination of the ellipse in the elliptic shape presence / absence determination process By storing the angle in the estimated circumference calculation unit 26, it is possible to perform only comparison of gradients with respect to the target coordinate in the estimated circumference area, as in the case of determining whether there is a perfect circle shape.

  Next, a third embodiment of the present invention will be described. FIG. 17 is a block diagram functionally illustrating members around the circular shape presence / absence determination unit 8 in the image processing apparatus of the present embodiment. As shown in FIG. 17, this embodiment has the same configuration as the image processing apparatus 101 except that it includes a circumference determination threshold value changing unit 27 that is a circumference determination threshold value changing unit. Has common functions.

  The circumference determination threshold value changing unit 27 changes the circumference determination threshold value. The circumference determination threshold value can be changed by an instruction from the user using an input unit (not shown). Moreover, you may be comprised so that a circumference determination threshold value may be changed on predetermined conditions irrespective of a user's instruction | indication. For example, the determination threshold is changed according to the reading resolution of the image reading apparatus connected to the image processing apparatus. As described above, the present embodiment includes the circumference determination threshold value changing unit 27 that can change the circumference determination threshold value, so that even if the image reading apparatus deteriorates over time, the determination is performed with a certain accuracy or more. be able to.

  Next, an image processing apparatus 103 according to the fourth embodiment of the present invention will be described. FIG. 18 is a block diagram functionally showing the configuration of the image processing apparatus 103. As shown in FIG. 18, the image processing apparatus 103 has the same configuration as that of the image processing apparatus 101 except that the image processing apparatus 103 includes a filter unit 28 that is a filter unit, and common members have common functions. Yes.

  The filter unit 28 includes an R filter unit 28R, a G filter unit 28G, and a B filter unit 28B.

  The R filter unit 28R reads out the R signal value from among the RGB signal values of the input image preprocessed by the preprocessing unit 1 and performs filtering. For example, the R filter unit 28R extracts only pixels having an R signal value of 200 to 220 in 256 gradations. The extracted pixel information is output to a processing unit (not shown) of the filter unit 28. The G filter unit 28G reads out the G signal value from among the RGB signal values of the input image preprocessed by the preprocessing unit 1 and performs filtering. For example, the G filter unit 28G extracts only pixels having an R signal value of 30 to 40 in 256 gradations. The extracted pixel information is output to the processing means of the filter unit 28. The B filter unit 28B reads out the B signal value from among the RGB signal values of the input image preprocessed by the preprocessing unit 1 and performs filtering. For example, the B filter unit 28B extracts only pixels having a B signal value of 10 to 20 in 256 gradations. The extracted pixel information is output to the processing means of the filter unit 28.

  The processing means in the filter unit 28 extracts pixels of a specific color using the pixel information output from the R filter unit 28R, the G filter unit 28G, and the B filter unit 28B. For example, in the above example, the specific color is a color having an RGB signal value of (200 to 220, 30 to 40, 10 to 20), and pixels of this specific color are extracted. Pixels other than the specific color that are not extracted are pixels whose RGB signal values are (0, 0, 0). The image after extraction of the specific color is output to the preprocessing unit 1 and stored in the preprocessing unit 1. Thereafter, the image processing apparatus 103 performs circular shape presence / absence determination processing in the same manner as the image processing apparatus 101 using the image after extraction of the specific color as an input image.

  As described above, the image processing apparatus 103 can determine the presence / absence of a circular shape using the extracted image of a specific color as an input image. Therefore, when the circular color distribution to be determined is known in advance, for example, an imprint When the color is vermilion, the filter unit 28 can cut an image of a color other than vermilion, and the determination accuracy can be improved. Further, even if the input image is a color image, the image processing apparatus 103 can perform the determination with the color image without performing the achromatization process, and therefore can perform the determination with high accuracy. Furthermore, since the image after the specific color extraction has less information than the original image, the storage means of the image processing apparatus 103 can be reduced in capacity.

  Next, an image processing apparatus 104 according to the fifth embodiment of the present invention will be described. FIG. 19 is a block diagram functionally showing the configuration of the image processing apparatus 104. As shown in FIG. 19, the image processing apparatus 104 has the same configuration as the image processing apparatus 103 except that it includes a binarization processing unit 29 that is a binarization processing unit, and common members are common. It has a function.

  The binarization processing unit 29 performs binarization processing on the image after the specific color extraction. For example, the pixel of the specific color is “1”, and all the RGB signal values are (255, 255, 255). Pixels other than the specific color are set to “0” and all RGB signal values are set to (0, 0, 0). The binarized image is output to the preprocessing unit 1 and stored in the preprocessing unit 1. Thereafter, the image processing apparatus 104 performs a circular shape presence / absence determination process in the same manner as the image processing apparatus 101 using the image after extraction of the specific color as an input image.

  As described above, the image processing apparatus 104 can determine the presence or absence of a circular shape using the binarized image as an input image. Since the image after the binarization process has less information than the original image, the storage means of the image processing apparatus 104 can be reduced in capacity.

  Next, an image processing system 200 that is an embodiment of the image processing system of the present invention will be described. FIG. 20 is a block diagram functionally showing the configuration of the image processing system 200. As illustrated in FIG. 20, the image processing system 200 includes an image reading device 9, an image processing device 101, and a notification device 30. Since the image reading device 9 and the image processing device 101 are the reading device and the processing device, respectively, the details are omitted.

  An image read from the document image by the image reading device 9 is output to the image processing device 101, and the image processing device 101 uses the image as an input image to determine whether or not the input image has a circular shape. When it is determined that a circumferential image exists in the input image, the image processing apparatus 101 outputs the circumferential presence determination, the estimated circle center, and the circle existence possibility area to the notification device 30. When it is determined that no circumference image exists in the input image, the image processing apparatus 101 outputs a circumference absence determination to the notification apparatus 30.

  The notification device 30 notifies the user based on the information output from the image processing device 101. The notification may be sound from a speaker (not shown), or a sentence indicating the determination result may be displayed on a display unit (not shown). For example, when the circumference existence determination, the estimated circle center, and the circle existence possibility region are output from the image processing apparatus 101, the notification device 30 inputs the information to the display unit together with the announcement “circumference image detected”. An image and a circle existence possibility area may be displayed. Further, for example, when the circumference nonexistence determination is output from the image processing apparatus 101, the notification apparatus 30 may make an announcement “No circumference image was detected”. Thus, the image processing system 200 can notify the determination result to the user using at least one of sound and light.

  The image processing method of the present invention is realized by causing the first to fifth embodiments of the image processing apparatus and the embodiment of the image processing system to execute a circular presence / absence determination process. As an embodiment of the present invention relating to a program, an image processing program for performing circular shape presence / absence determination processing may be recorded on a computer-readable recording medium in which a program to be executed by a computer is recorded.

  As a result of recording such an image processing program on a recording medium, a recording medium on which a program code (execution format program, intermediate code program, source program) for executing circular shape presence / absence determination processing is recorded can be provided in a portable manner. .

  The recording medium is read by a digital color image forming apparatus or a program reading apparatus provided in a computer system, thereby executing the image processing method.

  The computer system includes an image input device such as a flatbed scanner, a film scanner, and a digital camera, a computer that performs various processes such as the image processing method by loading a predetermined program code, and a CRT that displays the processing results of the computer. An image display device such as a display and a liquid crystal display, and a printer that outputs the processing results of the computer to paper or the like. Further, the computer system is provided with a network card, a modem, and the like as communication means for connecting to a server or the like via a network.

  In the present embodiment, as the recording medium, a memory (not shown) such as a ROM (Read Only Memory) itself may be a program medium because processing is performed by a microcomputer, or an external storage device (not shown). As a program medium that can be read by inserting a recording medium therein.

  In any case, the stored program may be configured to be accessed and executed by the microprocessor, or in any case, the program code is read and the read program code is stored in the microcomputer. It may be downloaded to a program storage area (not shown) and the program may be executed. It is assumed that this download program is stored in the main device in advance.

Here, the program medium is a recording medium configured to be separable from the main body, and includes a tape system such as a magnetic tape and a cassette tape, a magnetic disk such as a flexible disk and a hard disk, and a CD-ROM (Compact Disc Read Only Memory). / MO (Magneto Optical
disc) / MD (Mini Disc) / DVD (Digital Versatile Disc) optical discs, IC (Integrated Circuit) cards (including memory cards) / optical cards, etc., mask ROM, EPROM (Erasable Programmable) It may be a medium that carries a fixed program code including a read only memory (EEPROM), an electrically erasable programmable read only memory (EEPROM), and a semiconductor memory such as a flash ROM.

  In addition, as long as the system configuration can connect to a communication network including the Internet, a medium that fluidly carries the program code so as to download the program code from the communication network may be used. When the program is downloaded from the communication network in this way, the download program may be stored in the main device in advance or may be installed from another recording medium. The present invention can also be realized in the form of a computer data signal embedded in a carrier wave in which the program code is embodied by electronic transmission.

2 is a block diagram functionally showing the configuration of an image processing apparatus 101. FIG. 3 is a block diagram functionally showing the configuration of an estimated value calculation unit 4. FIG. 3 is a block diagram functionally showing a configuration of an ideal gradient calculation unit 5. FIG. 3 is a block diagram functionally showing the configuration of an approximate gradient calculation unit 6. FIG. 3 is a block diagram functionally showing the configuration of a circular shape presence / absence determining unit 8. FIG. It is a figure for demonstrating the characteristic of the template for perfect circle determination. It is a figure for demonstrating calculation of the feature-value by the feature-value calculation part 10. FIG. It is a figure for demonstrating calculation of the feature-value by the feature-value calculation part 10. FIG. It is a figure for demonstrating calculation of the feature-value by the feature-value calculation part 10. FIG. It is a figure for demonstrating estimation circle center calculation and circle presence possibility area | region calculation. It is a figure for demonstrating extraction of the specific edge pixel by the specific edge extraction part 20. FIG. It is a figure for demonstrating calculation of the difference vector by the difference vector calculation part 22. FIG. It is a figure which represents typically the comparison with the ideal gradient and approximate gradient by the gradient comparison part. It is a flowchart showing a perfect circle shape existence determination process. It is a figure for demonstrating the effect of the determination accuracy improvement in the image processing apparatus. It is a figure which shows the ellipse determination template. 2 is a block diagram functionally showing the configuration of an image processing apparatus 102. FIG. It is a figure for demonstrating an estimated circumference area | region. It is a block diagram functionally showing members around the circular shape presence / absence determining unit 8. 2 is a block diagram functionally showing the configuration of an image processing apparatus 103. FIG. 2 is a block diagram functionally showing the configuration of an image processing apparatus 104. FIG. 2 is a block diagram functionally showing the configuration of an image processing system 200. FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Pre-processing part 2 Imprint memory | storage part 3 Template feature memory | storage part 4 Estimated value calculation part 5 Ideal gradient calculation part 6 Approximate gradient calculation part 7 Gradient comparison part 8 Circular shape presence determination part 9 Image reader 10 Feature-value calculation part 11 Template comparison History storage unit 12 Feature amount comparison unit 13 Feature amount comparison result storage unit 14 Match determination unit 15 Estimated circle center calculation unit 16 Circle existence possibility region calculation unit 17 Difference vector calculation unit 18 Ideal tangent vector calculation unit 19 Gradient calculation unit 20 Identification Edge extraction unit 21 Specific edge storage unit 22 Difference vector calculation unit 23 Gradient calculation unit 24 Gradient comparison result storage unit 25 Circumference determination unit 101 to 104 Image processing device 200 Image processing system

Claims (12)

Template information storage means for storing information of the template image;
Estimated circle center calculation means for calculating an estimated circle center, which is a coordinate estimated that the center of the circle exists in the input image, using the information;
A circle existence possibility area calculating means for calculating, using the information, a circle existence possibility area that is an area in which the circle may exist in the input image;
An ideal gradient calculating means for calculating an ideal gradient, which is a gradient of a tangent of a circle passing through the target coordinate, with the estimated circular center as the center of the circle from the estimated circle center and the target coordinate in the circle existence possibility region When,
An approximate tangent gradient at the target coordinate of the specific image is calculated from the target coordinate in the specific image in the circle existence possibility region and the coordinate in the specific image and in the vicinity of the target coordinate. An approximate gradient calculating means for calculating an approximate gradient;
A gradient comparison means for comparing the ideal gradient and the approximate gradient having the same target coordinates used for calculating the gradient and outputting a gradient comparison result;
An image processing apparatus comprising: a circular shape presence / absence determining unit that determines whether or not a circumferential image is included in the input image based on the gradient comparison result. An estimated circumference calculating means for calculating an estimated circumference area that is a circumference area around the estimated circle center and that is an area within the circle existence possibility area;
The ideal gradient calculating means calculates an ideal gradient that is a gradient of a tangent of a circle passing through the target coordinate from the estimated circle center and the coordinate of interest in the estimated circumferential region, with the estimated circle center as the center of the circle. Calculate
The approximate gradient calculating means approximates the specific image at the target coordinate from the target coordinate in the specific image within the estimated circumferential area and the coordinate in the specific image and in the vicinity of the target coordinate. The image processing apparatus according to claim 1, wherein an approximate gradient that is a gradient of a typical tangent is calculated. The approximate gradient calculating unit includes a specific edge extracting unit that extracts a specific edge pixel that is a specific edge pixel in the circle existence possibility region, a specific edge storage unit that stores the coordinates of the specific edge pixel, and the specific unit First difference calculation means for calculating a coordinate difference between edge pixels,
The image processing apparatus according to claim 1, wherein the specific image is an image including the specific edge pixels, and a target coordinate used for calculating an approximate gradient is selected from the specific edge pixels.   The ideal gradient calculation means includes a second difference calculation means for calculating a coordinate difference between the coordinates in the circle existence possibility area and the estimated circle center, and converts the coordinate difference to set the estimated circle center as the center of the circle. The image processing apparatus according to claim 1, further comprising: an ideal tangent vector calculation unit that calculates a tangent vector of a circumference that passes through coordinates in a circle existence possibility region.   The circular shape presence / absence determining means determines whether or not a circumferential image is included in the input image based on a gradient comparison result storage means for storing the gradient comparison result, and the gradient comparison result and a circumference determination threshold value. 5. The image processing apparatus according to claim 1, further comprising: a circumference determination unit that determines   The image processing apparatus according to claim 5, further comprising a circumference determination threshold value changing unit that changes the circumference determination threshold value. Filter means for extracting an image of a specific color from the input image;
2. An image having a specific color extracted by the filter means as an input image, and configured to be able to determine whether or not a circumferential image is included in the input image. The image processing apparatus as described in any one of -6. Binarization processing means for performing binarization processing using information on the image of the specific color and information on the image other than the specific color;
An image after binarization processing by the binarization processing means is used as an input image, and it is configured to be able to determine whether or not a circumferential image is included in the input image. The image processing apparatus according to claim 7.   An image processing apparatus according to claim 1, and a notification apparatus that notifies a user of a result of determination by the image processing apparatus using at least one of sound and light. Image processing system. An estimated circle center calculation step of calculating an estimated circle center, which is a coordinate estimated that the center of the circle exists in the input image, using the information of the template image;
A circle existence possibility area calculating step for calculating a circle existence possibility area that is an area in which the circle may exist in the input image using the information;
An ideal gradient calculation step of calculating an ideal gradient that is a gradient of a tangent of a circle passing through the coordinate of interest, with the estimated circle center being the center of the circle from the estimated circle center and the coordinate of interest in the circle existence possibility region When,
An approximate tangent gradient at the target coordinate of the specific image is calculated from the target coordinate in the specific image in the circle existence possibility region and the coordinate in the specific image and in the vicinity of the target coordinate. An approximate gradient calculating step for calculating an approximate gradient;
A gradient comparison step of comparing the ideal gradient and the approximate gradient having the same target coordinates used for calculating the gradient and outputting a gradient comparison result;
And a circular shape presence / absence determination step for determining whether or not a circumferential image is included in the input image based on the gradient comparison result. Computer
Template information storage means for storing information of the template image;
Estimated circle center calculation means for calculating an estimated circle center, which is a coordinate estimated that the center of the circle exists in the input image, using the information;
A circle existence possibility area calculating means for calculating, using the information, a circle existence possibility area that is an area in which the circle may exist in the input image;
An ideal gradient calculating means for calculating an ideal gradient, which is a gradient of a tangent of a circle passing through the target coordinate, with the estimated circular center as the center of the circle from the estimated circle center and the target coordinate in the circle existence possibility region When,
An approximate tangent gradient at the target coordinate of the specific image is calculated from the target coordinate in the specific image in the circle existence possibility region and the coordinate in the specific image and in the vicinity of the target coordinate. An approximate gradient calculating means for calculating an approximate gradient;
A gradient comparison means for comparing the ideal gradient and the approximate gradient having the same target coordinates used for calculating the gradient and outputting a gradient comparison result;
An image processing program that functions as circular shape presence / absence determining means for determining whether or not a circumferential image is included in the input image based on the gradient comparison result.   A computer-readable recording medium on which the program according to claim 11 is recorded.

Images