JP4206604B2 - Pattern detection apparatus and method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a pattern detection apparatus and method for detecting a specific pattern included in an image.
[0002]
[Prior art]
Conventionally, for example, in one of the data processing performed in a copying machine, the acquired image data is stored in the image memory as binary data or multi-value data, and each pixel of the reference pattern stored in the memory in advance is stored with this data. There is known a pattern matching process for detecting a specific pattern included in an image, recognizing a shape, and the like by collating with the above data.
[0003]
By the way, in recent years, as an input device for scanning and an output device for outputting image data as a print print, those capable of processing image data having a large amount of information such as high resolution data at high speed are becoming widespread. In order to realize data processing including pattern matching processing in real time in cooperation with these input / output devices, there is a problem that the hardware configuration is not limited. In order to solve this problem, there is a need for a pattern detection apparatus that can detect a specific pattern at high speed and that has a simple configuration.
[0004]
[Problems to be solved by the invention]
As such a pattern detection device, for example, after the binary data is reduced in resolution, (1) an ON pixel (that is, a dot is present in a rectangular block region of m × n pixels around the target pixel). The number of existing pixels) falls within a predetermined range, and (2) a predetermined shape and size as a part of a specific pattern are provided based on the condition that there are no ON pixels in the predetermined area around the block area. Those that recognize partial images are known. However, in this pattern detection device, it is not possible to sufficiently specify a partial image having a hollow shape in particular, and other images (for example, a circular image with a solid interior or the like) There has been a problem of erroneously detecting a quadrangle or triangular pixel image having a size of 1).
[0005]
The present invention has been made in view of the above technical problem, and can detect a specific pattern with high accuracy and high speed, and has a simple configuration and a specific pattern using the pattern detection device. An object is to provide a detection method.
[0006]
[Means for Solving the Problems]
  The invention according to claim 1 of the present application is binarized by a binarization unit that binarizes image data and a binarization processing unit in a pattern detection apparatus that detects a specific pattern included in an image. Partial image recognition means for recognizing a partial image having a hollow shape that is a part of the specific pattern included in the image data, and the partial image recognition meansIsScan each block area in the image data with a size that fits the partial imageShi,Based on a predetermined condition, the recognition target included in the block area is evaluated to detect the partial image candidate, the size of the recognition target detected as the partial image candidate is detected, and the size of the recognition target is determined. In the corresponding pixel block whose size is variably set,Reference block consisting of a pixel of interest and pixels in a predetermined range around itSet,In the above reference blockThe recognition target that satisfies the condition that at least one off pixel exists is determined as a partial image.
[0007]
  The invention according to claim 2 of the present application is the invention according to claim 1,The partial image recognition meansthe aboveThe size of the recognition target, that is, the shape and dimensions of the recognition target, is detected as the size of a rectangular pixel block that contains the recognition target.It is characterized by that.
[0008]
  Furthermore, the invention according to claim 3 of the present application is a pattern detection method for detecting a specific pattern included in an image, the step of binarizing image data, and the image binarized in the binarization step. A partial image recognition step for recognizing a partial image having a hollow shape that is part of the specific pattern included in the data, and the partial image recognition stepIs, While scanning each block area in the image data having a size that can accommodate the partial image,A sub-step of detecting a recognition target included in the block region based on a predetermined condition and detecting a candidate of the partial image; a sub-step of detecting a size of the recognition target detected as the partial image candidate; In the pixel block in which the size is variably set corresponding to the size of the recognition target,Reference block consisting of a pixel of interest and pixels in a predetermined range around itSub-steps to determine,In the above reference blockA recognition target that satisfies the condition that at least one off pixel exists is determined as a partial image.Including sub-steps,It is characterized by that.
[0009]
  Furthermore, the invention according to claim 4 of the present application is a computer-readable recording medium on which a pattern detection program used for detecting a specific pattern included in an image is recorded, and the pattern detection program , A step of binarizing image data, and a partial image for recognizing a partial image having a hollow shape that is part of the specific pattern included in the image data binarized in the binarization step A partial image recognition step.Is, While scanning each block area in the image data having a size that can accommodate the partial image,A sub-step of detecting a recognition target included in the block region based on a predetermined condition and detecting a candidate of the partial image; a sub-step of detecting a size of the recognition target detected as the partial image candidate; In the pixel block in which the size is variably set corresponding to the size of the recognition target,Reference block consisting of a pixel of interest and pixels in a predetermined range around itSub-steps to determine,In the above reference blockA recognition target that satisfies the condition that at least one off pixel exists is determined as a partial image.Including sub-steps,It is characterized by that.
[0010]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.
FIG. 1 and FIG. 2 are a diagram schematically showing a copier according to Embodiment 1 of the present invention, and a block diagram showing a basic configuration of the copier. The copying machine 1 basically includes an image scanning unit 4 that reads an original using an optical system and acquires image data, and an image data processing unit that performs various processes including character recognition processing on the image data. 5 and a printing unit 6 that executes printing based on the image data. As can be seen from FIG. 2, these components include a CPU 11 that controls various components incorporated in the copying machine 1 via the data bus 15, a ROM 12 that stores a control program for the CPU 11, and a CPU 11. It is connected to a RAM 13 that temporarily stores programs and data to be executed for control.
[0011]
In this copying machine 1, first, in the image scanning unit 4, a document on the document table is optically read by the CCD 2 and subjected to photoelectric conversion. As the CCD 2, a three-line linear CCD corresponding to each color component of R (red), G (green), and B (blue) is used. When irradiating light on a document on a document table, this linear CCD can acquire R, G, and B color light from the reflected light in one scan, and each color light is converted into an electrical signal (analog image data). ). The analog image data acquired by the image scanning unit 4 is input to the image data processing unit 5 and processed so as to have an output format suitable for the printing unit 6. The printing unit 6 creates a print print based on the image data output from the image data processing unit 5. The operation of each component during this time is executed by the program recorded in the ROM 12 being stored in the RAM 13 and sequentially read out by the CPU 12.
[0012]
The image data processing unit 5 incorporates a pattern detection processing unit that detects a pattern having a predetermined shape and size printed on a document. FIG. 3 is a block diagram showing the configuration of the pattern detection processing unit incorporated in the image data processing unit 5. The pattern detection processing unit 20 includes an image input unit 21 to which digitized image data is input, a resolution reduction processing unit 22 that reduces the resolution of the image data, and each of the image data that has been reduced in resolution. And a binarization processing unit 23 that performs labeling (binarization) after determining the density value of the pixel.
[0013]
The image data input to the image input unit 21 is data in which each color of R (red), G (green), and B (blue) has a density value of 8 bits (256 gradations). In the image input unit 21, preprocessing such as resolution conversion and scaling is applied to the image data as necessary. The binarization processing unit 22 determines whether the density value for a predetermined color (pattern color) of each pixel constituting the image data is included in a preset “reference density range”; A binarization process is performed so that a pixel having a density value within the reference density range is set as an on pixel. The “on pixel” means a pixel having a value 1 in the binary image data, that is, a pixel having a point, and the same applies to the following description. Thereafter, the image data is input to the resolution reduction processing unit 23 and thinned to be converted into low resolution data to be processed in the subsequent element recognition process. The reduced-resolution image data is then temporarily stored in a predetermined memory (for example, RAM 13).
[0014]
As can be seen from FIG. 3, the pattern detection processing unit 20 recognizes an element that is a part of a specific pattern included in the image data, and determines a pattern composed of a plurality of elements. It has a pattern determination processing unit 25 to perform and a score setting processing unit 26 for setting a score to the determination result.
The element recognition processing unit 24 reads out image data stored in the RAM 13 and recognizes an element in the image data. The “element” is a partial image constituting a part of the pattern, and is an image having a hollow shape, size, and color. The result of the element recognition process is stored in the RAM 13 again. Based on the arrangement of the plurality of elements stored in the RAM 13, the pattern determination processing unit 25 determines whether or not a specific pattern is configured from these elements. And the score setting process part 26 sets the score according to it with respect to the determination result by the pattern determination process part 25. FIG.
[0015]
FIG. 4 is a flowchart of pattern detection processing by the pattern detection processing unit 20 having the above-described configuration. In this pattern detection process, first, the color image data input via the image input unit 21 is binarized with respect to the color of the pattern (S11). Next, the binary image data is converted into lower resolution image data (S12). Further, the entire binary image data with a reduced resolution is sequentially scanned with a filter of a size (m × n pixel block) in which the elements can be accommodated, and the elements in the image data are recognized (S13). This element recognition process will be described later in detail with reference to FIG. Subsequently, the binary image data is scanned with a filter having a size in which a specific pattern can be accommodated, and the pattern is determined based on the arrangement of recognized elements (S14). Then, the acquired pattern determination result is evaluated, and a score corresponding to it is set (S15). Above, a pattern detection process is complete | finished.
[0016]
FIG. 5 is a block diagram showing the configuration of the element recognition processing unit 24 in the pattern detection processing unit 20. The element recognition processing unit 24 evaluates the recognition target in the image data based on a predetermined condition, and determines the recognition target satisfying the condition as an “element candidate”. An element candidate size detection unit 32 that detects the size of the recognized recognition target, and an off-pixel that is present in a reference block including a pixel of interest in a pixel block region including the recognition target and pixels in a predetermined range around the pixel of interest. It includes an off-pixel detection unit 33 and a final element determination unit 34 that receives final determination and detection results from these components and performs final element determination. In this embodiment, reading or storing of data performed between the element recognition processing unit 24 and the RAM 13 is controlled via the memory control unit 18.
Note that the “target pixel” refers to the center of each pixel block area or the center of each pixel block area when the pixel block area in the image data is sequentially scanned by a filter of a size (m × n pixel block) in which the element is accommodated in the element recognition process. A reference pixel in the vicinity thereof is represented, and the “off pixel” represents a pixel having a value of 0 in the binary image data, that is, a pixel having no point. The same applies to the following description.
[0017]
During the element recognition process, the memory control unit 18 calculates an address on the RAM 13 in which the binary image data is stored, and sequentially reads pixel block regions in the binary image data from the RAM 13. This pixel block area is a rectangular block area composed of m × n pixels centered on the pixel of interest, and its size is determined in relation to the processing resolution and the ideal element size. The pixel block area read by the memory control unit 18 is input to the element candidate determination unit 31 and the off pixel detection unit 33.
[0018]
The element candidate determination unit 31 evaluates a recognition target in the input pixel block region based on a predetermined condition, and determines a recognition target that satisfies the condition as an “element candidate”. The above predetermined conditions are: (1) all the pixels constituting the outermost line of the pixel block area are off pixels, and (2) the on-state included in the predetermined area centered on the pixel of interest in the pixel block area. The number of pixels falls within a specified range.
[0019]
For example, as shown in FIG. 6, ideal elements are arranged symmetrically on the top and bottom and left and right, except for 2 × 2 pixels at the center and 4 pixels at the top in the 4 × 4 pixel block area. This is a hollow circular image. In FIG. 6, the hatched pixels are on pixels. With respect to this ideal element, element candidate determination in the 7 × 7 pixel block region is performed. In this case, as conditions corresponding to the above conditions (1) and (2), all the pixels constituting the outermost line of the 7 × 7 pixel block area are off pixels, and within the 7 × 7 pixel block area It is set that the number of ON pixels included in the 5 × 5 pixel block region centering on the target pixel is within the range of 6 to 10. As a result of evaluation based on such conditions, recognition targets included in various 7 × 7 pixel block areas as shown in FIG. 7 are both determined as element candidates. Although these recognition targets differ from the ideal elements as shown in FIG. 6 in terms of shape or size, they satisfy the above conditions (1) and (2), and are thus determined as element candidates. On the other hand, recognition targets included in various 7 × 7 pixel block areas as shown in FIG. 8 do not satisfy at least one of the above conditions (1) and (2), and are not determined as element candidates.
A pixel block region including a recognition target determined as an element candidate by the element candidate determination unit 31 is input to the element candidate size detection unit 32 and the final element determination unit 34 as needed.
[0020]
The element candidate size detection unit 32 receives a pixel block region including a recognition target determined as an element candidate by the element candidate determination unit 31 and detects the size of the recognition target. At this time, the size of the recognition target is expressed by using the size of a rectangular pixel block that accommodates the recognition target (that is, an m × n pixel block). For example, when the recognition target is the same as an ideal element as shown in FIG. 6, the size is represented as “4 × 4 pixel block”. The detection result by the element candidate size detection unit 32 is input to the off pixel detection unit 33 as needed.
[0021]
The off-pixel detection unit 33 receives the detection result from the element candidate size detection unit 32 and, based on the detection result, first corresponds to the size of the recognition target from the pixel block region input from the RAM 13 via the memory control unit 18. A pixel block having a size to be extracted is extracted. Subsequently, in the extracted pixel block, a reference block including a target pixel included in the extracted pixel block and pixels in a predetermined range around the target pixel is set, and an off pixel in the reference block is detected. When an off pixel is detected in the reference block, the recognition target included in the pixel block is determined to be an element having a hollow shape.
[0022]
FIGS. 9A to 9H show setting examples of reference blocks in various pixel blocks extracted as pixel blocks corresponding to the size of the recognition target in the off-pixel detection process. In the various pixel blocks shown in the figure, a pixel with a star mark represents a target pixel, and a hatched area represents a reference block.
In the 3 × 3 pixel block 41 shown in FIG. 9A, the target pixel is the center pixel, and only the center pixel is set as the reference block. When the target pixel is an off pixel, the recognition target is determined to be an element having a predetermined shape. Further, in the 3 × 4 pixel blocks 42 and 43 shown in FIG. 9B, the target pixel is one of the two pixels inside the block, and the target pixel and one pixel inside the block adjacent to the target pixel are used. A reference block to be configured is set. In the 3 × 4 pixel blocks 42 and 43 in which the position of the target pixel is different from each other, the reference block is set in the same area inside the block. The same applies to the 4 × 3 pixel blocks 44 and 45 shown in FIG.
[0023]
Further, in the 3 × 5 pixel block 46 shown in FIG. 9D, the pixel of interest is a central pixel, and a reference block composed of the central pixel and two pixels inside the adjacent block is set. ing. Furthermore, in the 4 × 4 pixel block 47, 48, 49, 50 shown in FIG. 9E, the target pixel is any one of the four pixels in the block, and the target pixel and A reference block composed of three pixels inside an adjacent block is set.
[0024]
Further, in the 5 × 3 pixel block 51 shown in FIG. 9F, the pixel of interest is the central pixel, as in the 3 × 5 pixel block 46 shown in FIG. A reference block composed of two pixels inside the block adjacent to it is set. Further, in the 4 × 5 pixel blocks 52 and 53 shown in (g) of FIG. 9, the target pixel is either one of the two pixels near the center inside the block, and the target pixel and 5 in the block adjacent to the target pixel are included. A reference block composed of pixels is set. The same applies to the 5 × 4 pixel block shown in FIG.
[0025]
Thus, the size of the reference block is variable corresponding to the size of the extracted pixel block, that is, the size of the recognition target determined as an element candidate by the element candidate determination unit 31. That is, various processing resolutions and element sizes can be handled.
The detection result by the off-pixel detection unit 33 is sent to the final element determination unit 34 at any time.
[0026]
The final element determination unit 34 receives the determination result from the element candidate determination unit 31 and the detection result from the off-pixel detection unit 33, and finally determines the recognition target included in the pixel block region as an element based on the determination result. To do. The determination result is sent to and stored in the RAM 13 via the memory control unit 18.
[0027]
Through the above processing, recognition targets that cannot satisfy all of the various conditions are eliminated. This makes it possible to select a recognition target that is more suitable for an ideal element as shown in FIG. 6, and particularly to recognize an element having a hollow shape with high accuracy. Further, in this embodiment, since the element recognition processing as described above is performed after the resolution of the image data is reduced, the processing speed can be increased and the memory used can be reduced. Furthermore, in this embodiment, since the element is a substantially hollow circular image, the shape to be recognized can be specified under relatively simple conditions as a shape that is vertically and horizontally symmetrical.
[0028]
FIG. 10 is a flowchart of the element recognition process (S13 in FIG. 4). In this process, for example, image data is scanned with a filter of a size that accommodates the ideal element shown in FIG. 6, and it is determined whether each pixel block region in the filter satisfies the determination condition as described above. . This flowchart describes the case where a filter corresponding to a 7 × 7 pixel block region is used.
In this process, first, binary image data stored after the resolution is reduced is read (S21). Next, the scanning start position of the filter is determined. That is, the Y-direction coordinate of the filter is set to be positioned at the upper end of the image data (S22), and the X-direction coordinate of the filter is set to be positioned at the left end of the image data (S23).
[0029]
When scanning is started, it is first determined whether or not all the outermost lines of the 7 × 7 pixel block region are off pixels in the 7 × 7 pixel block region in the filter (S24). If it is determined that all pixels are not off, it is determined that the recognition target in the pixel block area is not an element, and the element flag is set to “off” (S29). Then, it progresses to S30. On the other hand, if all are off pixels, the process proceeds to S25. In S25, it is determined whether or not the number of ON pixels in the 5 × 5 pixel block region is 6 to 10. As a result, if it is not 6 to 10, the element flag is set to “off” (S29), and the process proceeds to S30. On the other hand, if it is 6 to 10, the process proceeds to S26.
[0030]
In S26, the size of the element candidate included in the 7 × 7 pixel block area is checked. In S27, it is determined whether or not an off pixel exists in the reference block set in the 5 × 5 pixel block corresponding to the size of the element candidate. As a result, if there is no off pixel, the element flag is set to “off” (S29), and the process proceeds to S30. On the other hand, if there is an off pixel, the process proceeds to S28. In S28, since all the conditions of S24, S25, and S27 are satisfied, it is determined that the recognition target in the pixel block region is an element having a predetermined shape, and the element flag is set to “ON”. Then, it progresses to S30.
[0031]
In S30, it is determined whether or not the 7 × 7 pixel block region to be recognized is the final pixel block region in the X direction. As a result, if it is not the final pixel block region, the filter is shifted by one pixel in the X direction (S31), and the process returns to S24 and the subsequent steps are performed. On the other hand, if it is the final pixel block region, the process proceeds to S32, and it is determined whether or not the 7 × 7 pixel block region to be recognized is the final pixel block region in the Y direction. As a result, if it is not the final pixel block region, the filter is shifted by one pixel in the Y direction (S33), the process returns to S23, and the subsequent steps are performed. On the other hand, if it is the final pixel block area, the element recognition process is terminated and the process returns to the main routine.
[0032]
Such element recognition processing in S21 to S33 is executed based on an operating program incorporated in the copying machine 1, and in this embodiment, the program is stored in the ROM 12 in the copying machine 1 (FIG. 2). (See). Without being limited thereto, such a program may be stored in a file format on an external recording medium such as a floppy disk or a CD-ROM.
[0033]
Note that the present invention is not limited to the illustrated embodiments, and it goes without saying that various improvements and design changes are possible without departing from the scope of the present invention. For example, each component of the element recognition processing unit 24 described with reference to FIG. In this case, the pixel block region is simultaneously input from the RAM 13 to each component (the element candidate determination unit 31, the element candidate size detection unit 32, and the off pixel detection unit 33), and then the element candidate determination unit 31 and the off pixel detection. The determination and detection results by the unit 33 are individually input to the final element determination unit 34. The final element determination unit 34 comprehensively evaluates the input determination and detection result, and determines whether or not the recognition target is an element having a predetermined shape.
[0034]
【The invention's effect】
  As is clear from the above description, according to the invention according to claim 1 of the present application, in the pattern detection device that detects a specific pattern included in an image, binarization means for binarizing image data; Partial image recognition means for recognizing a partial image having a hollow shape that is a part of the specific pattern included in the image data binarized by the binarization processing means. Image recognition meansIsScan each block area in the image data with a size that fits the partial imageShi,Based on a predetermined condition, the recognition target included in the block area is evaluated to detect the partial image candidate, the size of the recognition target detected as the partial image candidate is detected, and the size of the recognition target is determined. In the corresponding pixel block whose size is variably set,Reference block consisting of a pixel of interest and pixels in a predetermined range around itSet,In the above reference blockSince the recognition target that satisfies the condition that there is at least one off pixel is determined as a partial image, the recognition target that further matches the ideal partial image is selected, and the partial image having a hollow shape is recognized with high accuracy. CanThe size of the reference block is variable corresponding to the size of the recognition target included in the block area in the image data, and various processing resolutions and element sizes can be handled.
[0035]
  According to the invention of claim 2 of the present application, the aboveThe size of the recognition target is detected as the size of a rectangular pixel block that contains the recognition target..
[0036]
  Furthermore, according to the invention according to claim 3 of the present application, in the pattern detection method for detecting a specific pattern included in an image, the image data is binarized in the binarization step and the binarization step. A partial image recognition step for recognizing a partial image having a hollow shape that is a part of the specific pattern included in the image data, and the partial image recognition stepIs, While scanning each block area in the image data having a size that can accommodate the partial image,A sub-step of detecting a recognition target included in the block region based on a predetermined condition and detecting a candidate of the partial image; a sub-step of detecting a size of the recognition target detected as the partial image candidate; In the pixel block in which the size is variably set corresponding to the size of the recognition target,Reference block consisting of a pixel of interest and pixels in a predetermined range around itSub-steps to determine,In the above reference blockA recognition target that satisfies the condition that at least one off pixel exists is determined as a partial image.Including sub-steps,Therefore, it is possible to select a recognition object that is more suitable for an ideal partial image, and to recognize a partial image having a hollow shape with high accuracy.
[0037]
  Furthermore, according to the invention according to claim 4 of the present application, there is provided a computer-readable recording medium recording a pattern detection program used when detecting a specific pattern included in an image, the pattern detection A program recognizes a step of binarizing image data and a partial image having a hollow shape that is a part of the specific pattern included in the image data binarized in the binarization step. A partial image recognition step, and the partial image recognition stepIs, While scanning each block area in the image data having a size that can accommodate the partial image,A sub-step of detecting a recognition target included in the block region based on a predetermined condition and detecting a candidate of the partial image; a sub-step of detecting a size of the recognition target detected as the partial image candidate; In the pixel block in which the size is variably set corresponding to the size of the recognition target,Reference block consisting of a pixel of interest and pixels in a predetermined range around itSub-steps to determine,In the above reference blockA recognition target that satisfies the condition that at least one off pixel exists is determined as a partial image.Including sub-steps,Therefore, it is possible to select a recognition object that is more suitable for an ideal partial image, and to recognize a partial image having a hollow shape with high accuracy.
[Brief description of the drawings]
FIG. 1 is a diagram schematically showing a copying machine according to an embodiment of the present invention.
FIG. 2 is a block diagram showing a basic configuration of the copying machine.
3 is a block diagram showing a configuration of a pattern detection processing unit in the image data processing unit of FIG. 2;
FIG. 4 is a flowchart of pattern detection processing.
FIG. 5 is a block diagram showing a configuration of an element recognition processing unit in the pattern detection processing unit.
FIG. 6 is a diagram illustrating an arrangement of pixels constituting an ideal element.
FIG. 7 is a diagram showing various 7 × 7 pixel block areas including a recognition target determined as an element candidate.
FIG. 8 is a diagram illustrating various 7 × 7 pixel block regions including recognition targets that are not determined as element candidates.
FIG. 9 is a diagram illustrating a setting example of a reference block in a pixel block including element candidates of various sizes.
FIG. 10 is a flowchart for element recognition processing;
[Explanation of symbols]
1 ... Copier
5. Image data processing unit
13 ... RAM
18 ... Memory control unit
20 ... pattern detection processing unit
21. Image input unit
22: Binarization processing unit
23. Resolution reduction processing unit
24 ... Element recognition processing unit
25. Pattern determination processing unit
26 ... Score setting processing section
31 ... Element determination part
32 ... Element candidate size detection unit
33: Off-pixel detection unit
34 ... Final element determination section

Claims (4)

In a pattern detection apparatus for detecting a specific pattern included in an image,
Binarization means for binarizing image data;
Partial image recognition means for recognizing a partial image having a hollow shape that is part of the specific pattern included in the image data binarized by the binarization processing means,
The partial image recognition means,
It scans each block region in the image data having a size of the partial image fits,
Evaluating a recognition target included in the block area based on a predetermined condition to detect a candidate for the partial image;
Detect the size of the recognition target detected as a candidate for the partial image,
In a pixel block whose size is variably set corresponding to the size of the recognition target, a reference block including a pixel of interest and pixels in a predetermined range around it is determined .
A pattern detection apparatus that determines a recognition target that satisfies a condition that at least one off pixel exists in the reference block as a partial image. The pattern detection apparatus according to claim 1 , wherein the partial image recognition unit detects the size of the recognition target, that is, the shape and dimensions of the recognition target, as a size of a rectangular pixel block in which the recognition target is accommodated . In a pattern detection method for detecting a specific pattern included in an image,
Binarizing image data; and
A partial image recognition step for recognizing a partial image having a hollow shape that is a part of the specific pattern included in the image data binarized in the binarization step,
The partial image recognition step includes
While scanning each block area in the image data having a size that can accommodate the partial image,
A sub-step of evaluating a recognition target included in the block region based on a predetermined condition to detect a candidate for the partial image;
A sub-step of detecting a size of a recognition target detected as a candidate for the partial image;
A sub-step for defining a reference block including a pixel of interest and pixels in a predetermined range around the pixel in a pixel block whose size is variably set corresponding to the size of the recognition target ;
And a sub-step of determining, as a partial image, a recognition target that satisfies a condition that at least one off pixel exists in the reference block . A computer-readable recording medium recording a pattern detection program used when detecting a specific pattern included in an image, the pattern detection program comprising:
Binarizing image data; and
A partial image recognition step for recognizing a partial image having a hollow shape that is a part of the specific pattern included in the image data binarized in the binarization step,
The partial image recognition step includes
While scanning each block area in the image data having a size that can accommodate the partial image,
A sub-step of evaluating a recognition target included in the block region based on a predetermined condition to detect a candidate for the partial image;
A sub-step of detecting a size of a recognition target detected as a candidate for the partial image;
A sub-step for defining a reference block including a pixel of interest and pixels in a predetermined range around the pixel in a pixel block whose size is variably set corresponding to the size of the recognition target ;
A computer-readable recording medium comprising: a sub-step of determining a recognition target satisfying a condition that at least one off pixel exists in the reference block as a partial image.

Images