JP2018088206A - Circle detection method, circle detection device, program and storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a circle detection method, a device, a program thereof and a storage medium capable of detecting a circle in a short time.SOLUTION: A circle detection method 1 comprises a binarization step generating a binarization image in a detection target area and a detection target outside area, a labeling step labeling each binarized detection target area, a contour extraction step extracting a contour of a figure for each binarized detection target area, and a circle area securing step in which for each detection of a contour pixel forming a contour by scanning in extracted contour data, the contour pixel is regarded as a point located on a circumference of a circle, a pixel position in a predetermined direction detecting a presence of a circle and separated by a length corresponding to a radius of a predetermined circle desired to be detected is regarded as a center of the circle desired to be detected, if a plurality of coordinates in predetermined positions suitable for checking whether they are the same label within a range surrounded by the radius are all the same label, determining as a circle, and if a center of the circle desired to be detected is not present within the already secured circle area, securing as an area of a new circle.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a circle detection method, a circle detection device, a circle detection program, and a storage medium storing a circle detection program in an image among image processing techniques.

  Patent Document 1 discloses an image processing method for detecting a center position of each circle with respect to a plurality of circles including a circle having a size different from that of other circles displayed as image data. A great circle region identifying step for identifying a region including the largest circle, which is the largest circle, and an edge pixel in the great circle region, and a large circle candidate pixel that constitutes the circumference of the great circle by detecting the edge pixel in the great circle region A large circle candidate pixel detecting step for detecting the large circle candidate pixel, a large circle detecting step for detecting the center position and radius of the great circle by performing multi-stage circle approximation on the detected large circle candidate pixel, and Labeling within the great circle so as not to include the great circle edge based on the center position and radius, a small circle region identifying step for identifying one or more regions including one small circle, and detecting edge pixels in the small circle region The detected edge pixel is composed of a small circle. A small circle candidate pixel detecting step for detecting as a small circle candidate point pixel to be detected, and a small circle for performing a multi-stage circle approximation on the detected small circle candidate pixel to obtain a center position and a radius of the small circle An image processing method having a detection step is disclosed.

  Patent Document 2 discloses an image processing method for classifying an instrument by processing an image of an instrument photographed by an image photographing means, an image input step for inputting an image of the instrument, and an input of the input instrument Edges are extracted from the image, binarization processing is performed on the image from which the edges have been extracted, thinning processing is performed on the binarized image, and the contour line of the display panel of the instrument is extracted A contour extracting step, a virtual contour creating step for creating a circular virtual contour using Hough transform based on the extracted contour, and a contour extracted by the contour extracting step When the degree of coincidence between the image of the contour line after the expansion processing and the image of the virtual contour line created by the virtual contour line creating step is larger than a predetermined threshold value, the instrument The circle An instrument classification step for classifying into an instrument, wherein the contour extraction step is performed on the instrument image input in the image input step when the instrument classification step does not classify the instrument as a round instrument. A background removal process for changing an image outside the display panel area of the instrument to a black pixel is performed, and a smoothing process and a binarization process are sequentially performed on the image of the display panel area after the background removal process. The region having the largest area is selected as the largest region among the connected components of the graphic pixels in the image subjected to the value processing, and the contour line of the selected largest region is extracted as the contour line of the display panel of the instrument. The virtual contour line creating step expands the contour line image extracted by the contour line extracting step, and from the contour line image subjected to the expansion processing, a quadrangular shape that is the shape of the display board A contour line is extracted as a virtual contour line, and the instrument classification step includes an image of the contour line after the expansion process and a quadrangular contour line created by the virtual contour line creation step and having a shape of the display panel An image processing method for classifying the instrument into a square instrument when the degree of coincidence with the image is greater than a predetermined threshold is disclosed.

  Patent Document 3 discloses an image recognition device including a circle detection unit that detects a circle from a captured image obtained by capturing an image of the front of the vehicle with an imaging device mounted on a vehicle, and includes the circle detection unit among the captured images. A ratio that is at least one of an inner area of an inscribed rectangle inscribed in the circle detected in step 1 and a predetermined area between the inscribed rectangle and the circumscribed rectangle inscribed on the circle detected by the circle detection unit An image recognizing device including a sign recognizing unit for recognizing a circular sign based on a color ratio of a determination area is disclosed.

JP 2004-220370 A Japanese Patent No. 3799407 Japanese Patent No. 5370375

  In the invention described in Patent Document 1, when detecting the center position and radius of the great circle, multi-stage circle approximation is performed on the large circle candidate pixels, and when the center position and radius of the small circle are further detected. It is a technology that performs multi-stage circle approximation on small circle candidate pixels and detects the center position and radius of the circle, and is not a technology that detects a circle quickly in the screen, so there are many In the case where a circle exists, there is a problem that the processing time for detecting a circle in an image becomes longer by further obtaining the center position and radius of the circle after the circle is detected.

  The invention described in Patent Document 2 has a problem that it takes a lot of processing time to search for a circle in an image because it uses thinning processing, expansion processing, and Hough transform.

  As described in paragraphs [0028] and [0029] of cited document 3, the invention of Patent Document 3 uses a Hough transform for circle detection, and therefore requires a long processing time for searching for a circle in an image. There was a problem.

  The present invention has been devised in view of these problems, and a circle detection method, a circle detection device, a program thereof, and a memory capable of detecting a circle in an extremely short time without using the Hough transform generally used for circle detection It is an object to provide a medium.

  A circle detection method 1 according to claim 1 is a circle detection method 1 for detecting a circle in an image 3 based on image data, and a detection target region 4 and a non-detection target region 5 with respect to the image 3. A binarization step 11 for generating the binarized image, a labeling step 12 for labeling each binarized detection target region 4, and the figure for each binarized detection target region 4 In the contour extraction step 13 for extracting the contour 9, and in the extracted contour data, the y-axis direction 8 is detected from one of the x-axis directions 7 for each pixel of the y-axis of the contour data or for each pixel of the x-axis. Each time a contour pixel 6 that scans from one side to form a contour 9 is detected, the contour pixel 6 is positioned at a circumferential portion of the circle, and is detected in a predetermined direction for detecting the presence of a circle and a predetermined target to be detected. The pixel position separated by the length of the radius of the circle is It is determined as a circle when a plurality of coordinates at a predetermined position 40, which are suitable for confirming whether or not they are the same label within the range surrounded by the radius, are the same label as the center of the circle to be extracted, and the detection And a circle area securing step 14 for securing a new circle area when the center 41 of the desired circle does not exist within the already secured circle area.

  In the circle detection method 1 according to claim 2, in the case of detecting a plurality of size circles according to claim 1, the size of the radius is sequentially reduced from the maximum value among the plurality of circles to minimize the circle. The circle area securing step 14 is repeatedly performed up to a value, and the sequentially reduced circles are determined, and the sequentially reduced circle areas are secured.

  The circle detection device 2 according to claim 3 is a circle detection device 2 that detects a circle in the image 3 based on image data, and the detection target region 4 and the non-detection target region 5 with respect to the image 3. A binarizing unit that generates a binarized image, a labeling unit that performs labeling for each of the binarized detection target regions 4, and an outline of the figure for each of the binarized detection target regions 4 In the extracted contour data, scanning is performed from one of the x-axis directions 7 for each pixel of the y-axis or one of the y-axis directions 8 for each pixel of the x-axis in the extracted contour data. Each time the contour pixel 6 forming the contour 9 is detected, the contour pixel 6 is positioned at the circumferential portion of the circle, and the radius of the predetermined circle to be detected in a predetermined direction for detecting the presence of the circle is detected. The circle where the pixel position separated by the length of The center of the circle to be detected is determined as a circle when a plurality of coordinates at a predetermined position 40 suitable for checking whether or not the same label is within the range surrounded by the radius are the same label. And a circle area securing means for securing 41 as a new circle area when 41 does not exist within the already secured circle area.

  When detecting a plurality of size circles, the circle detection device 2 according to claim 4 is the minimum of the plurality of circles by sequentially reducing the size of the radius from the maximum value. The circle area securing means is repeatedly performed up to a value, a circle that is sequentially reduced is determined, and an area of the circle that is sequentially reduced is secured.

  The program according to claim 5 is a program for causing a computer to execute the circle detection method 1 according to claim 1 or 2.

  A storage medium according to a sixth aspect is a computer-readable storage medium storing the program according to the fifth aspect.

The present invention has an effect that the processing time for detecting a circle from the image 3 can be remarkably shortened. Comparing the number of searches from the screen 3 until the circle detection is completed between the case of the Hough transform and the case of the present invention, the circle detection by the Hough transform is performed at the pixel position after detecting the edge. In order to search for a circle by drawing a circle having a predetermined radius r as the center (pixel number 2πr) for every pixel on the circumference (pixel number 2πr), the search number is (2πr). ² ) In contrast, in the case of the present invention, the detection is performed at the pixel position after edge detection, and the pixel position is set to one position on the circumference of the circle within a predetermined radius r in a predetermined direction. Since only a predetermined number of pixel positions are searched, the number of searches is (predetermined number of directions × predetermined number of search positions), so the number of searches can be greatly reduced, the circle detection processing speed is increased, and circle detection is performed. Has the effect of significantly reducing processing time .

  Further, in the present invention, when performing circle detection when overlapping circles exist, the circle having the size desired to be detected is detected by sequentially detecting the smallest circle from the largest circle among the size of the circle to be obtained. Even if it overlaps, there exists an effect that all can be detected.

It is a flowchart of the circle detection method of this invention. It is a block diagram which shows the structure of the circle | round | yen detection apparatus of this invention. It is explanatory drawing explaining the detection object area | region in an image, and a non-detection object area | region. It is an image figure which shows the circle | round | yen which exists in an image. It is a binarization figure which shows the binarization of an image in a binarization process. It is label data explanatory drawing of the detection object area | region after binarization by a labeling process. It is outline data explanatory drawing which extracted the outline of the detection target area | region after binarization by the outline extraction process. It is explanatory drawing of outline data, label data, and area ensuring data when detecting three circles. It is explanatory drawing of the detection of a great circle in a circle area ensuring process. It is explanatory drawing of the outline data after extracting a great circle, label data, and area | region securing data. It is explanatory drawing of the detection of the small circle after securing a great circle. It is explanatory drawing of the outline data after detecting a small circle after a great circle detection, label data, and area | region securing data. It is explanatory drawing of the outline data after detecting three circles, label data, and area | region securing data. It is explanatory drawing of the outline data, label data, and area ensuring data when detecting one circle. It is explanatory drawing explaining the scanning of a contour pixel search for one circle. It is explanatory drawing of the circular area ensuring process of performing circle detection based on the outline pixel of one circle, (a) is explanatory drawing of a search direction, (b) is explanatory drawing of a search area range. It is explanatory drawing of the circle detection of the range of an upper direction based on the outline pixel of one circle. It is explanatory drawing of the circle detection of the range of the right direction based on the outline pixel of one circle. It is explanatory drawing of the circle detection of the range of a downward direction based on the outline pixel of one circle. It is explanatory drawing of the outline data, label data, and area | region ensuring data after detecting one circle.

  A circle detection method 1 according to the present invention is a circle detection method 1 for detecting a circle in an image 3 based on image data, as shown in FIG. A binarization step 11 for generating a binarized image of the non-target region 5, a labeling step 12 for labeling each binarized detection target region 4, and the binarized detection target region 4 In the contour extraction step 13 for extracting the contour 9 of the figure every time, and in the extracted contour data, y for each pixel of the y-axis of the contour data from one of the x-axis directions 7 or y for each pixel of the x-axis Each time the contour pixel 6 that scans from one of the axial directions 8 to form the contour 9 is detected, the contour pixel 6 is located at a circumferential portion of the circle in a predetermined direction for detecting the presence of the circle and Pixels separated by the length of the radius of the predetermined circle you want to detect If the position is the center of the circle to be detected and the coordinates of the predetermined position 40 suitable for confirming whether or not the same label is within the range surrounded by the radius are all the same label, it is determined as a circle. And a circle area securing step 14 for securing a new circle area when the center 41 of the circle to be detected does not exist within the already secured circle area.

  The contour extraction process 13 stores the contour data, the labeling process 12 stores the label data, and the circle area securing process 14 stores the area securing data.

  The image data is acquired in an image acquisition step 10 for acquiring image data by photographing or image reading.

  First, the binarization step 11 will be described. The detection target region 4 and the non-detection target region 5 are, for example, when only the circle 31 is detected when a figure of a circle 31, a triangle 32, and a quadrangle 33 is present in the image data in FIG. Uses the circle 31 as the detection target region 4, the base 30, the triangle 32, and the quadrangle 33 as the non-detection target region 5, and binarizes the detection target region 4 and the non-detection target region 5. For example, as shown in FIG. 4, when only the circle is set as the detection target region 4, a single small circle figure and a figure in which both the small circle and the great circle partially overlap and become the detection target region 4 are used. The binarization in which the background 30 portion becomes the non-detection target region 5 is performed, and a binarized image as shown in FIG. 5 is created.

  Next, a labeling step 12 and a contour extracting step 13 are performed on the binarized image 3. The labeling process 12 will be described. As shown in the binarized image 3, when two detection target areas 4 are detected, A and B are labeled for each area, and label data A and B associated with coordinates are stored. . When the stored label data is shown in the figure, it is stored as shown in the label data explanatory diagram 61 of FIGS.

  The contour extraction process 13 will be described. When two detection target areas 4 are detected as shown in the binarized image 3, the contour is detected for each area. Contour data associated with coordinates is stored. When the stored contour data is shown in the figure, it is stored as shown in the contour data explanatory diagram 60 of FIGS.

  Next, the circle area securing step 14 will be described. A predetermined direction for detecting the presence of a circle is set based on the radius of a predetermined circle to be detected in advance or a specific pixel position (contour pixel position) 6. The radius of the predetermined circle may be one radius, a plurality of radii, or all possible radii of the circle to be obtained in the image 3. The predetermined direction for detecting the presence of a circle based on a specific pixel position (outline pixel position) 6 is, for example, left and right in the x-axis direction 7 at 90 ° intervals as shown in FIG. 16 (a) or (b). 4 directions of 90 ° intervals in the vertical direction and y-axis direction 8 may be set, 8 directions of 45 ° intervals, 12 directions of 30 ° intervals, and different angles may be used. An interval of It is desirable that the set interval be constant during the circle search.

  In the contour data, as shown in FIG. 15, the contour 9 is scanned from one side in the x-axis direction 7 for each pixel on the y-axis of the contour data or from one side in the y-axis direction 8 for each pixel on the x-axis. Are detected.

  In the detected contour pixel 6, the pixel separated from the contour pixel 6 in the predetermined direction for detecting the presence of the circle and the radius of the predetermined circle to be detected, with the contour pixel 6 positioned at the circumferential portion of the circle The position is the center 41 of the circle to be detected, and it is confirmed whether or not a plurality of pixels at a predetermined position 40 suitable for confirming whether or not the same label is arbitrarily set within the range surrounded by the radius are the same label. To do. Since the predetermined position 40 can be determined as to whether or not it is a circle, a suitable number and position for confirmation are set. For example, in FIG. 17 to FIG. 19, six predetermined positions 40 are set. However, several predetermined positions may be set in the vicinity of the circumferential edge of the circle. You may set several positions.

  Here, in order to explain one method of securing a circle area in an easy-to-understand manner, a case where one circle exists in the image 3 is taken up and a case where the area of the circle is used as a label C is shown based on FIGS. explain. As the data for entering circle detection, contour data and label data are obtained, but area securing data is not yet obtained. This is shown in FIG.

  First, using the contour data, for example, when the contour pixel 6 on the contour 9 is detected by scanning the screen 3 in the x-axis direction 7 as shown in FIG. 15, as shown in FIG. As the points where the contour pixels 6 are located in the circumferential portion of the circle, the predetermined directions for detecting the presence of the circle are four directions at intervals of 90 °, and as shown in FIG. The pixel position separated by the length of the radius is set as the center 41 of the circle to be detected, and the range 50 surrounded by the upper radius, the range 51 surrounded by the right radius, and the lower radius. It is determined whether each range of the range 52 and the range 53 surrounded by the radius in the left direction is a desired circle.

  For example, as shown in FIG. 17, the determination is made by checking whether all the pixels at the predetermined positions 40 of the six marks in the range 50 surrounded by the radius in the upward direction are all the same label, It is confirmed whether or not the center 41 of the circle to be detected does not exist within the already secured circle area. In the case of FIG. 17, since all the pixels at the predetermined position 40 are outside the range of the label C area, they are not determined as circles.

  Next, in the same contour pixel 6, as shown in FIG. 18, the pixel position separated by the length of the radius of a predetermined circle to be detected to the right is set as the center 41 of the circle to be detected, and the dotted line in FIG. In the range 51 surrounded by the radius shown in (5), for example, it is confirmed whether all the pixels at the predetermined positions 40 of the six crosses are the same label, and the center 41 of the circle to be detected is already secured. Check if it is not within the circled area. In the case of FIG. 18, one pixel is a label C, but the other five pixels are not determined as circles because the pixel at a predetermined position 40 is outside the range of the label area.

  Next, in the same outline pixel 6, as shown in FIG. 19, the pixel position separated by the length of the radius of a predetermined circle to be detected downward is set as the center 41 of the circle to be detected, and the dotted line in FIG. In the range surrounded by the radius shown (in FIG. 15, the dotted line overlaps with the solid line), for example, it is confirmed whether all the pixels at the predetermined positions 40 of the six X marks are all the same label. . In the case of FIG. 19, since the coordinates of all the predetermined positions 40 are the label C, it is determined as a circle.

  Then, when the circle center 41 to be detected does not exist in the already secured circle area, it is secured as a new circle area, and the circle center 41 to be detected exists in the already secured circle area. Do not secure as a circle area. In the case of FIG. 19, this corresponds to the case where the determined circle center 41 does not exist within the area of the already secured circle. Therefore, a new circle area is secured in the area securing data, and the circle is identified. The center and radius of the secured circle are determined, and the center, radius (diameter) and label are stored as area securing data. Symbols for identifying the circle are shown in the figure as shown in the area securing data explanatory diagram 62 shown in FIG.

  Next, every time the image 3 is further scanned and the contour pixel 6 is detected in the contour data, the predetermined range of four directions at 90 ° intervals, the range 50 surrounded by the upper radius, and the right radius It is determined whether each of a range 51 surrounded by a circle, a range 52 surrounded by a lower radius, and a range 53 surrounded by a left radius is a circle to be obtained. If the center 41 of the circle to be detected does not exist within the area of the already secured circle, the area is secured as a circle area, and if the center 41 of the circle to be detected exists within the area of the already secured circle, A method of not securing the area is performed by scanning all the pixels in the screen 3. The method for detecting one circle has been described above.

  Next, a case where a plurality of circles having different radii are overlapped will be described. For example, as shown in FIGS. 4 to 13, in the case of an image 3 in which a small circle and a part of a large circle overlap in addition to one small circle, the detection target region 4 is used. In this case, the method for securing the circle area is to sequentially reduce the size of the radius from the maximum value among the plurality of circles, repeat the circle area securing step 14 to the minimum value, and determine the sequentially reduced circle. The area of the circle that is gradually reduced is secured. That is, the circle having the largest radius within the conceivable range is detected, and the circle having the smallest radius is sequentially detected.

  As the data for entering circle detection, contour data and label data are obtained, but area securing data is not yet obtained. This is shown in FIG. Even when a circle having a set radius is not detected at the start of detection, existing circles are detected by sequentially reducing the radius. As shown in FIG. 9, the circle is detected by, for example, checking whether all the pixels at the predetermined positions 40 of the six X marks are the same label, and the center 41 of the circle to be detected is already secured. 9, in the case of FIG. 9, since the six pixels are labels A and the center 41 of the circle does not exist in the already secured circle area, the area is secured as a circle. do. Then, in the area securing data, the circle identification symbol A1, the center of the circle and the radius (diameter) are stored in association with the coordinates. This is shown in FIG. 62, which is an area reservation data explanatory diagram of FIG.

  Then, after the large circle shown in FIG. 10 is secured, the circle area securing step 14 is repeated in the small circle search to perform circle determination and circle securing. For example, as shown in FIG. 11, in the contour pixel 6 on the contour data, the pixel position separated by the length of the radius of a predetermined circle to be detected below is set as the center 41 of the circle to be detected, and the dotted line in FIG. For example, in the range surrounded by the radius shown in FIG. 11 (most of the dotted lines overlap the solid line in FIG. 11), for example, all the pixels at the predetermined positions 40 of the seven X marks are all the same label. To check. In the case of FIG. 11, since all the pixels are label data and label A, it is determined as a circle.

  Since this corresponds to the case where the center 41 of the circle to be detected does not exist within the area of the already secured circle, it is secured as a new circle area. Then, the newly secured small circle is set as the identification symbol A2, the center and radius (diameter) of the small circle are determined, and the center, radius (diameter) and identification symbol A2 of the small circle are determined in relation to the coordinates. Store as area reservation data. This is illustrated in the area securing data explanatory diagram 62 of FIG.

  Next, it is possible to complete the detection of all the circles in the image 3 by similarly performing the small circle of the label B. Thereby, the identification symbol B of the label B, the center of the small circle, and the radius (diameter) are stored as area securing data in association with the coordinates. This is shown in an area securing data explanatory diagram 62 in FIG.

  Since the number of searches from the screen 3 until the circle detection is completed can be greatly reduced by the present invention compared to the case of the Hough transform, the circle detection processing time from the image 3 can be significantly shortened.

  In the circle area securing step 14, when the radius of the circle to be obtained is not known but all circles in the image 3 are to be obtained, a maximum value and a minimum value considered as the radius are set in advance, and the minimum value is determined from the radius of the maximum value. The circular area securing step 14 is repeatedly performed at each radius. When circle determination and circle reservation are completed for all radii, detection of all circles in the image 3 is completed.

  Next, as shown in FIG. 2, the circle detection device 2 of the present invention includes an image processing unit 55 including a birectification unit 21, a labeling unit 22, a contour extraction unit 23, and a circular area securing unit 24, and an image acquisition unit. 20, a storage unit 25, a display unit 56, an operation unit 27, and a power supply unit 29. An imaging device 26 and a display device 56 are connected to the circle detection device 2.

  The image acquisition unit 20 acquires image data obtained by imaging the subject 57 with the imaging device 26 such as a camera, or image data read from the storage unit 25 in which the image data is stored. The display unit 28 displays the image 3 processed by the image processing unit 55 on a display device 56 such as a personal computer.

  Next, the image processing unit 55 of the circle detection device 2 of the present invention is means for performing circle detection processing for detecting a circle in the image 3 based on the image data. A binarization unit 21 that is a binarization unit that generates a binarized image of the region 4 and the non-detection target region 5, and a labeling unit that performs labeling for each of the binarized detection target regions 4 A contour extracting unit 23 which is a contour extracting means for extracting the contour 9 of the figure for each binarized detection target region 4; and in the extracted contour data, 1 of the y-axis of the contour data Each time the contour pixel 6 forming the contour 9 is detected by scanning from one of the x-axis directions 7 for each pixel or from one of the y-axis directions 8 for each pixel of the x-axis, the contour pixel 6 is circled. Detect the presence of a circle as a point located in the circumference A pixel position separated in a predetermined direction and corresponding to the length of the radius of a predetermined circle to be detected is set as the center of the circle to be detected, and a predetermined number for confirming whether or not there are a plurality of identical labels within the range surrounded by the radius If the pixels at position 40 are all the same label, it is determined as a circle, and if the center of the circle to be detected does not exist within the already secured circle area, a new circle area is secured. And a circular area securing unit 24 as means.

  Then, when detecting a plurality of size circles, the circle detection device 2 is the circle area securing unit that sequentially reduces the radius of the plurality of circles from the maximum value to the minimum value. The circle area securing unit 24 is repeatedly performed to determine a sequentially reduced circle and secure a sequentially reduced circle area.

  The present invention can also be realized by executing the following processing. A program capable of executing the circle detection method 1 by a computer, the program being supplied to a system or apparatus via a network or a storage medium, and a computer of the system or apparatus calling and executing the program .

  Also, a storage medium storing a program for causing a computer to execute the circle detection method 1, wherein the program stored in the storage medium storing the computer-readable program for realizing the circle detection method is called and executed It is. Examples of the storage medium include a CD-ROM, a hard disk, a flexible disk, a magnetic tape, and a magneto-optical disk.

1 circle detection method 2 circle detection device 3 image 4 detection target area 5 non-detection target area 6 contour pixel 7 x-axis direction 8 y-axis direction 9 contour 10 image acquisition process 11 binarization process 12 labeling process 13 contour extraction process 14 circle Area securing step 20 Image acquisition unit 21 Binarization unit 22 Labeling unit 23 Outline extraction unit 24 Circle region securing unit 25 Storage unit 26 Imaging device 27 Operation unit 28 Display unit 29 Power supply unit 30 Base 31 Circle 32 Triangle 33 Square 40 Predetermined Position 41 Center 50 Range 51 Range 52 Range 53 Range 55 Image processing unit 56 Display device 57 Subject 60 Outline data explanatory diagram 61 Label data explanatory diagram 62 Area reservation data explanatory diagram

A circle detection method 1 according to claim 1 is a circle detection method 1 for detecting a circle in an image 3 based on image data, and a detection target region 4 and a non-detection target region 5 with respect to the image 3. A binarization step 11 for generating a binarized image, a labeling step 12 for labeling each binarized detection target region 4, and the detection for each binarized detection target region 4 In the contour extraction step 13 for extracting the contour 9 of the graphic of the target area, and in the extracted contour data, y for each pixel on the y-axis of the contour data from one of the x-axis directions 7 or for each pixel on the x-axis Each time the contour pixel 6 that scans from one of the axial directions 8 to form the contour 9 is detected, the contour pixel 6 is located at a circumferential portion of the circle in a predetermined direction for detecting the presence of the circle and Separated by the length of the radius of the circle you want to detect If the prime position is the center of the circle to be detected, and a plurality of coordinates at a predetermined position 40 suitable for checking whether or not the same label is within the range surrounded by the radius are all determined to be a circle And a circle area securing step 14 for securing a new circle area when the center 41 of the circle to be detected does not exist within the already secured circle area.

The circle detection device 2 according to claim 3 is a circle detection device 2 that detects a circle in the image 3 based on image data, and the detection target region 4 and the non-detection target region 5 with respect to the image 3. Binarizing means for generating a binarized image, labeling means for labeling each binarized detection target area 4, and the detection target area for each binarized detection target area 4 A contour extracting means for extracting the contour 9 of the figure, and in the extracted contour data, the y-axis direction 8 from one of the x-axis directions 7 for each y-axis pixel of the contour data or for each pixel of the x-axis 8 Each time a contour pixel 6 forming a contour 9 by scanning from one side of the circle is detected, the contour pixel 6 is positioned at a circumferential portion of the circle in a predetermined direction for detecting the presence of a circle and a predetermined desired value to be detected. Before the pixel position separated by the length of the circle radius If the plurality of coordinates at a predetermined position 40, which are suitable for confirming whether or not they are the same label within the range surrounded by the radius, are the same label as the center of the circle to be detected, it is determined as a circle, and the detection And a circle area securing means for securing a new circle area when the center 41 of the desired circle does not exist within the already secured circle area.

Claims (6)

A circle detection method for detecting a circle in an image based on image data,
A binarization step for generating a binarized image of the detection target region and the non-detection target region for the image;
A labeling step of performing labeling for each of the binarized detection target regions;
A contour extracting step for extracting the contour of the figure for each binarized detection target region;
In the extracted contour data, a contour pixel that forms a contour is detected by scanning from one side in the x-axis direction for each pixel on the y-axis of the contour data or from one side in the y-axis direction for each pixel on the x-axis. Each of the circles to be detected is a pixel position separated in a predetermined direction for detecting the presence of the circle and corresponding to the length of the radius of the predetermined circle to be detected, with the contour pixel being positioned at the circumferential portion of the circle. The center of the circle to be detected is determined as a circle when a plurality of coordinates at a predetermined position suitable for checking whether or not they are the same label within the range surrounded by the radius are all the same label And a circle area securing step of securing a new circle area when the area does not exist within the already secured circle area.   When detecting a plurality of size circles, the size of the radius is sequentially reduced from the maximum value among the plurality of circles, and the circle area securing step is repeatedly performed to the minimum value, and the circles are sequentially reduced. The circle detection method according to claim 1, wherein the circle is determined and a sequentially reduced circle area is secured. A circle detection device for detecting a circle in an image based on image data,
Binarization means for generating a binarized image of a detection target area and a non-detection target area for the image;
Labeling means for performing labeling for each of the binarized detection target areas;
Contour extraction means for extracting the contour of the figure for each binarized detection target area;
In the extracted contour data, a contour pixel that forms a contour is detected by scanning from one side in the x-axis direction for each pixel on the y-axis of the contour data or from one side in the y-axis direction for each pixel on the x-axis. Each of the circles to be detected is a pixel position separated in a predetermined direction for detecting the presence of the circle and corresponding to the length of the radius of the predetermined circle to be detected, with the contour pixel being positioned at the circumferential portion of the circle. The center of the circle to be detected is determined as a circle when a plurality of coordinates at a predetermined position suitable for checking whether or not they are the same label within the range surrounded by the radius are all the same label And a circle area securing means for securing a new circle area when the area does not exist within the already secured circle area.   When detecting a plurality of sizes of circles, the size of the radius of the plurality of circles is sequentially reduced from the maximum value, and the circle area securing means is repeatedly performed to the minimum value, and the circles are sequentially reduced. The circle detection device according to claim 3, wherein the circle is determined and a sequentially reduced circle area is secured.   A program for causing a computer to execute the circle detection method according to claim 1.   A computer-readable storage medium storing the program according to claim 5.

Images