Contour detection method and contour detection system

Abstract

To provide a contour detection method and a contour detection system capable of detecting particles with less error.SOLUTION: The contour detection method is a contour detection method executed by a contour detection system that detects contours of particles that are stacked circular-shaped objects, and includes a measurement step (S1) of performing two dimensional measurement or three dimensional measurement of the particles, an obtaining step (S2) of obtaining a measurement result of shapes of the particles, and a determination step (S3 to S10) of identifying the particles based on shape feature values that are indices of the circular-shaped objects and pixel intensities that are indices of region identification of the particles in the measurement result, and detecting the contours of the particles.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] FIELD OF THE DISCLOSURE The present disclosure relates to a contour detection method and a contour detection system, particularly for detecting the contour of particles that are stacked circular shapes. [Background technology]

[0002] For example, in the ironmaking process, the coke in the blast furnace must have high strength and small particle size variation to ensure good gas permeability within the furnace. Increasing the density of the coal charged into the coke oven, for example, by using a molded coal charging method, is an effective method for improving coke strength. Molded coal is made by compressing and molding coal powder using methods such as briquetting to increase density. Using molded coal not only improves coke strength but also reduces particle size variation.

[0003] On the other hand, it is known that the particle size of the briquettes changes depending on the particle size of the main raw material, particle size of the auxiliary raw material, fluctuations in moisture or supply amount during the granulation stage, and the rotation speed and rotation angle of the granulator. When the particle size variation is large, action is taken to remove the produced briquettes whose particle size falls outside the standard range and reuse them. Therefore, technology that can identify briquettes and measure their particle size in real time is required.

[0004] Conventionally, particle identification methods generally use optical instruments. For example, a method is known in which an image is captured using a CCD camera and a particle size distribution is calculated based on the brightness transition (see Patent Documents 1 and 2), or a method is known in which three-dimensional measurement is performed and height information is used to identify stacked particles (see Patent Document 3). [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Laid-Open No. 2004-16983 [Patent Document 2] Japanese Patent Application Publication No. 8-210962 [Patent Document 3] Japanese Patent Application Publication No. 2019-174931 Summary of the Invention [Problem to be solved by the invention]

[0006] When identifying particles transported in a stacked state, it is necessary to detect at least the upper layer particles of the stacked particles. This is because, if the particle size of the upper layer particles is known, the particle size within the stacked layer can be estimated using known technology. However, the method of Patent Document 1 relies on the brightness of the image results obtained by two-dimensional measurement, making it difficult to distinguish between upper and lower layer particles, resulting in false detection of them as a single particle. Patent Document 2 considers shrinkage separation, circular separation, and wedge separation as image processing logic for identifying adjacent particles. However, shrinkage separation cannot separate particles when there is a large amount of overlap. Because the raw material is elliptical, circular separation cannot calculate the long and short sides of the ellipse based only on the area and center of gravity, making it impossible to properly identify particles. Furthermore, wedge separation has problems with processing speed. In addition, separation is not possible when the particles overlap in a way that results in only one depression. Furthermore, it is impossible to determine whether they are upper or lower particles, and since particles are identified by connecting opposing depressions with a straight line, it is impossible to properly capture grain boundaries. Furthermore, the method of Patent Document 3 uses height information from three-dimensional data, but the height distribution changes smoothly when the particle has a gently sloping circular surface. Therefore, the method of Patent Document 2 also does not have sufficient accuracy in distinguishing between upper and lower layer particles in a stack.

[0007] In view of the above circumstances, an object of the present disclosure is to provide a contour detection method and a contour detection system that are capable of detecting particles with little error. [Means for solving the problem]

[0008] (1) A contour detection method according to an embodiment of the present disclosure includes: A contour detection method performed by a contour detection system for detecting contours of stacked circular particles, comprising: The method includes a measurement step of performing two-dimensional or three-dimensional measurement of the particle, an acquisition step of acquiring measurement results of the particle shape, and a determination step of identifying the particle based on a shape feature amount, which is an index of a circular shape in the measurement results, and pixel intensity, which is an index of particle region identification, and detecting the outline of the particle.

[0009] (2) As one embodiment of the present disclosure, in (1), The measurement results are two-dimensional images or three-dimensional data.

[0010] (3) As an embodiment of the present disclosure, in (1) or (2), The determination step includes a process of emphasizing the boundaries of the particles and a process of separating stacked particles in which a plurality of the particles are stacked.

[0011] (4) As an embodiment of the present disclosure, in (3), The determining step performs a first adjacent particle separation on the measurement results after processing to emphasize the boundaries of the particles.

[0012] (5) As an embodiment of the present disclosure, in (4), The process of separating stacked particles includes performing a second adjacent particle separation on a target portion that is a candidate for a particle separated by the first adjacent particle separation if the target portion has one or more recesses.

[0013] (6) As an embodiment of the present disclosure, in any one of (1) to (5), The pixel intensity is brightness when the measurement result is a two-dimensional image, and height when the measurement result is three-dimensional data.

[0014] (7) As an embodiment of the present disclosure, in any one of (1) to (6), The shape feature amount is a shape coefficient that is an index that objectively indicates whether the shape is close to a circle or an ellipse.

[0015] (8) As an embodiment of the present disclosure, in any one of (1) to (7), The determination step includes: performing a first adjacent particle separation to separate stacked particles, each stacked with a plurality of adjacent particles, based on the pixel intensities; A first circularity determination is performed based on the shape feature amount to determine whether a first target portion, which is a candidate particle separated by the first adjacent particle separation, is a circular object.

[0016] (9) As an embodiment of the present disclosure, in (8), The determination step includes: A second circularity determination is performed to determine whether the first target portion is a circular object based on whether the first target portion has one or more recesses.

[0017] (10) As an embodiment of the present disclosure, in (9), The second circularity determination involves moving an inscribed circle inscribed in the contour of the first target portion along the contour, and identifying recesses in the first target portion based on the number of areas enclosed by the trajectory of the inscribed circle.

[0018] (11) As an embodiment of the present disclosure, in (9) or (10), The determination step includes: performing a second adjacent particle separation on the first target portion when the second circularity determination determines that the first target portion has one or more recesses; A third circularity determination is performed based on the shape feature amount to determine whether or not a second target portion, which is a candidate particle separated by the second adjacent particle separation, is a circular object.

[0019] (12) As an embodiment of the present disclosure, in (11), The determination step includes: The contours of the first target portion and the second target portion determined to be circular are output as contours of stacked particles.

[0020] (13) A contour detection system according to an embodiment of the present disclosure includes: 1. A contour detection system for detecting contours of stacked circular particles, comprising: a measuring device that performs two-dimensional or three-dimensional measurement of the particles; The contour detection device includes an acquisition unit that acquires measurement results of the shape of the particle, and a determination unit that identifies the particle based on a shape feature that is an index of a circular shape in the measurement results and pixel intensity that is an index of particle region identification, and detects the contour of the particle. [Effects of the Invention]

[0021] According to the present disclosure, it is possible to provide a contour detection method and a contour detection system that are capable of detecting particles with little error. [Brief explanation of the drawings]

[0022] [Figure 1] FIG. 1 is a diagram illustrating an example of the configuration of a contour detection system according to an embodiment of the present disclosure. [Figure 2] FIG. 2 is another diagram showing an example of the configuration of the contour detection system of FIG. [Figure 3] FIG. 3 is a flowchart illustrating the process of a contour detection method according to an embodiment of the present disclosure. [Figure 4] FIG. 4 is a diagram illustrating the recessed portion. [Figure 5] FIG. 5 is a diagram showing the results of contour detection in a comparative example in which particles are identified using a conventional contour detection method. [Figure 6] FIG. 6 is a diagram showing the results of contour detection in an example in which particles were identified using a contour detection method according to an embodiment of the present disclosure. [Figure 7A] FIG. 7A is a diagram for explaining the second circularity determination. [Figure 7B] FIG. 7B is a diagram for explaining the second circularity determination. [Figure 7C] FIG. 7C is a diagram for explaining the second circularity determination. [Figure 8A]FIG. 8A is a diagram for explaining the second circularity determination. [Figure 8B] FIG. 8B is a diagram for explaining the second circularity determination. [Figure 8C] FIG. 8C is a diagram for explaining the second circularity determination. [Figure 8D] FIG. 8D is a diagram for explaining the second circularity determination. DETAILED DESCRIPTION OF THE INVENTION

[0023] A contour detection method and a contour detection system 10 according to an embodiment of the present disclosure will be described below with reference to the drawings. Figures 1 and 2 are diagrams showing an example configuration of the contour detection system 10. Figure 1 is a block diagram including an example configuration of a measurement device 13 and a contour detection device 14 provided in the contour detection system 10. Figure 2 shows the overall configuration of the contour detection system 10.

[0024] The contour detection system 10 detects the contours of stacked particles 12. Each particle 12 is circular or elliptical in plan view. In other words, the particle 12 is a circular object. Here, "circular" or "elliptical" refers to a shape that is closer to a circle or ellipse overall than a polygonal shape, and the contour may be partially linear or partially missing. The particles 12 are not limited to a specific type, but in this embodiment, they are described as molded coal. Molded coal is produced by uniformly compressing raw materials such as coal and then transported in a stacked state. As shown in FIG. 2, the particles 12 (molded coal) are transported in a stacked state in the transport direction by a belt conveyor 11. The contour detection system 10 according to this embodiment measures the molded coal during transport and detects the contours of the stacked molded coal. However, conventional contour detection methods have difficulty distinguishing between upper and lower layer particles, and partially overlapping upper and lower layer particles may be erroneously detected as a single particle 12. The contour detection system 10 according to this embodiment performs the process of the contour detection method described below, thereby enabling detection of upper particles of the particles 12 with minimal error. Here, stacked particles 12 does not only refer to a state in which the layers of the particles 12 are clearly separated, but also refers to a state in which at least some of the particles 12 are stacked one on top of the other.

[0025] As described above, the contour detection system 10 includes a measuring device 13 and a contour detection device 14. The measuring device 13 performs two-dimensional or three-dimensional measurement of the particles 12. Two-dimensional measurement is measuring the two-dimensional shape (planar shape) of the target object (particles 12 in this embodiment). Three-dimensional measurement is measuring the three-dimensional shape (solid shape) of the target object. The measuring device 13 outputs the measurement results to the contour detection device 14. The measurement results are two-dimensional images or three-dimensional data. The three-dimensional data is data that indicates the measured shape using coordinates in the conveying direction, width direction, and height direction (the stacking direction of the conveyed particles 12). Here, the width direction is a direction perpendicular to the conveying direction and height direction. In this embodiment, the measuring device 13 is a CCD camera arranged at an inspection location where the produced molded coal is measured, measures the two-dimensional shape, and outputs a two-dimensional image as the measurement result. The CCD camera may be installed above the molded coal transported by the belt conveyor 11.

[0026] The contour detection device 14 includes an acquisition unit 15, a determination unit 16, and an output unit 17. The acquisition unit 15 acquires measurement results of the shape of the particles 12. The determination unit 16 identifies the particles 12 based on the shape feature values ​​and pixel intensities in the measurement results, and detects the contours of the particles 12. The output unit 17 outputs the contour detection results. Here, the contour detection device 14 may have a hardware configuration such as a computer. The computer may be a server computer or a portable computer such as a laptop or tablet. In this embodiment, the contour detection device 14 is a computer used at an inspection site for produced molded coal.

[0027] FIG. 3 is a flowchart illustrating the process of the contour detection method executed by the contour detection system 10 according to this embodiment.

[0028] The CCD camera, which is the measuring device 13, measures the molded coal, which is the particles 12 being conveyed (step S1, measuring step).

[0029] The acquisition unit 15 acquires the measurement results of the shape of the particles 12 from the measurement device 13 (step S2, acquisition step). In this embodiment, the measurement results are two-dimensional images of the particles 12.

[0030] The determination unit 16 executes a process for emphasizing the boundaries of the particles 12 (step S3). In this embodiment, the process for emphasizing the boundaries of the particles 12 is a process for binarizing the pixel intensity values ​​of a two-dimensional image. The pixel intensity is an index for identifying particle regions, and when a two-dimensional image is used, a specific example of this is brightness. However, the pixel intensity may be any information possessed by each pixel in the measurement results, and another specific example of this may include height (the position in the stacking direction of the transported particles 12). Specifically, when the measurement results are two-dimensional images, the pixel intensity is brightness, and when the measurement results are three-dimensional data, the pixel intensity is height. The process for emphasizing the boundaries of the particles 12 makes it possible to distinguish the transported particles 12 from other parts (e.g., equipment such as the belt conveyor 11). For example, since the particles 12 are illuminated by lighting installed in the molded coal inspection area, it is possible to distinguish parts brighter than a threshold as particles 12 and parts darker than the threshold as equipment.

[0031] Here, the determination unit 16 may perform a process to remove unevenness in pixel intensity of the two-dimensional image before the process to emphasize the boundaries of the particles 12. The unevenness in pixel intensity occurs, for example, due to the degree of illumination on the particles 12. The process to remove unevenness in pixel intensity may be performed on the entire two-dimensional image using a known method (shading correction, for example).

[0032] The determination unit 16 performs a process (adjacent particle separation) to separate adjacent particles 12 from the two-dimensional image that has undergone the process of enhancing the boundaries of the particles 12 (step S4). A known method may be used for adjacent particle separation, and in this embodiment, the watershed method is used. The watershed method creates a distance map between the particles 12 and the rest of the background, and scores pixels that are far from the background and pixels that are close to it. The scored two-dimensional image is then treated like a map with contour lines, and particle 12 regions are identified by analyzing the boundaries of each peak based on minute changes in pixel intensity. That is, the gradient of the brightness distribution is calculated, and minute changes in the valleys between the peaks (locations with high brightness) are used as the criterion for dividing the regions, thereby identifying the particle 12 regions. As another example, the level set method or active contour method may be used for adjacent particle separation.

[0033] The determination unit 16 performs circularity determination on a portion (target portion or first target portion) in the two-dimensional image that is a candidate for a particle 12 separated by adjacent particle separation using a structure factor (step S5, first circularity determination). Although a single particle 12 has a circular shape, because the particles 12 are stacked, a stacked particle formed by stacking multiple particles 12 may be recognized as a single particle 12. For example, a stacked particle may be recognized as a single particle 12 with a distorted shape because some of the particles are upper particles that constitute the upper layer and some of the particles are lower particles that constitute the lower layer, and because they overlap in a planar view. Therefore, the determination unit 16 determines whether the target portion is a single particle 12 or a stacked particle based on whether the shape is close to a circle or an ellipse. The shape feature used in the circularity determination is an index of a circular shape and is the following index that objectively indicates whether a shape is close to a circle or an ellipse. Hereinafter, circle includes an ellipse, and "circle" means "circle or ellipse."

[0034] In this embodiment, the shape feature is a shape coefficient calculated by "A x B - 1." A and B will be described later. The shape coefficient approaches 0 when the shape is close to a circle, and is 0 when the shape is circular. Therefore, the determination unit 16 sets a threshold value for the shape coefficient in advance, and can objectively determine that the target part is circular when the calculated shape coefficient is smaller than the threshold value. Also, when the calculated shape coefficient is equal to or greater than the threshold value, it can determine that the target part is not circular.

[0035] In the formula for calculating the shape factor, A is anisotropy. A is calculated as "Ra / Rb". Here, Ra is the radius of the major axis of the area identified as particle 12. Rb is the radius of the minor axis of the area identified as particle 12. Also, B is bulkiness. B is calculated as "π×Ra×Rb / S". Here, S is the area of ​​the target portion.

[0036] Here, the shape feature quantity is not limited to the above-mentioned shape coefficients, as long as it can numerically represent the degree of circularity. As another example, roundness may be used as the shape feature quantity. Roundness is defined in JIS B0621-1984, "Definition and Display of Geometric Deviation," and is the magnitude of deviation of a circular feature from a geometrically correct circle. Since roundness also approaches 0 when a shape is close to a circle, it can be treated in the same way as the above-mentioned shape coefficients.

[0037] If the shape coefficient calculated in the circularity determination is smaller than the threshold and the determination unit 16 determines that the target portion is circular (Yes in step S6), the determination unit 16 proceeds to the process of step S10. That is, the determination unit 16 determines that one particle 12 (upper layer particle) has been separated (see particle 12 at the left end of FIG. 4 ), and proceeds to the process of step S10 without executing the processes of steps S7 to S9 (processes for separating stacked particles and extracting upper layer particles). In the example of FIG. 3 , when it is determined that one particle 12 has been separated by circularity determination using shape features, the process for separating stacked particles and extracting upper layer particles is not executed. Therefore, the processing load can be reduced compared to when the process for separating stacked particles and extracting upper layer particles is executed for the entire two-dimensional image.

[0038] If the shape coefficient calculated in the circularity determination is equal to or greater than the threshold and the determination unit 16 determines that the target portion is not circular (No in step S6), the determination unit 16 proceeds to the processing of step S7. That is, the determination unit 16 determines that the stacked particles have been erroneously detected as a single particle 12 (see the particles 12 in the center and right end of FIG. 4), and proceeds to the processing of separating the stacked particles (adjacent particle separation). As shown in FIG. 4, in stacked particles, recesses occur at the portions where the boundaries of upper and lower particles overlap. The recesses are portions where the boundaries are recessed toward the inside of the particle 12. The determination unit 16 identifies recesses in the target portion (step S7). Here, identifying recesses in the target portion means identifying the presence or absence of recesses that require adjacent particle separation processing. Then, based on whether or not the target portion has one or more recesses, it is determined whether or not the target portion is a circular object (second circularity determination).

[0039] As shown in Figures 7A to 7C and 8A to 8D, the determination unit 16 moves an inscribed circle 20 inscribed in the outline of the particle 12 along the outline, and identifies recesses in the layered particle based on the number of regions 21 enclosed by the trajectory (trajectory line) of the inscribed circle 20. In the example of Figure 7A, which targets the particle 12, the determination unit 16 gradually increases the diameter of the inscribed circle 20 as shown in Figures 7B and 7C. In addition, in the example of Figure 8A, which targets the particle 12, the determination unit 16 gradually increases the diameter of the inscribed circle 20 as shown in Figures 8B, 8C, and 8D.

[0040] Figure 7A shows an example of a stacked particle in which two particles 12a and 12b overlap. As shown in Figure 7B, when the diameter of the inscribed circle 20 is small, the number of regions 21 enclosed by the trajectory line is one. As shown in Figure 7C, as the diameter of the moving inscribed circle 20 is gradually increased, the region 21 enclosed by the trajectory line is separated into two or more regions. In this case, the stacked particle in Figure 7A is determined to have a recess that requires adjacent particle separation processing. Here, the number of regions 21 enclosed by the trajectory line can be counted using known image processing techniques. The stacked particle in Figure 8A shows an example in which the degree of overlap between particles 12a and 12b is greater than in the example in Figure 7A, and there is no significant difference in characteristic values ​​such as particle size between the lower and upper particles, making adjacent particle separation processing unnecessary. As shown in Figures 8B and 8C, even when the diameter of the inscribed circle 20 is gradually increased, the number of regions 21 enclosed by the trajectory line remains one. As shown in Figure 8D, if the diameter of the inscribed circle 20 is further increased, the region 21 surrounded by the trajectory line disappears. In this case, it is determined that there are no recesses in the stacked particle of Figure 8A that require adjacent particle separation processing. Here, the method for identifying recesses is not limited to the method described with reference to Figures 7A to 7C and Figures 8A to 8D. For example, recesses may be identified by detecting changes in the curvature of the boundary (contour) of the target portion using a known method.

[0041] The determination unit 16 also counts the number of recesses in the target portion. If the target portion has one or more recesses (Yes in step S7), the determination unit 16 performs adjacent particle separation on the target portion (step S8). Here, the adjacent particle separation uses the watershed method, as in step S4. The adjacent particle separation in step S8 (second adjacent particle separation) may be performed using the same processing as the adjacent particle separation in step S4 (first adjacent particle separation), but may be performed more efficiently by using a setting value obtained in the adjacent particle separation in step S4. For example, the determination unit 16 may set a threshold value for determining contours based on the gradient of brightness change in the target portion that is smaller than the threshold value used in step S4. That is, the determination unit 16 may set conditions using the setting value obtained in the adjacent particle separation in step S4 to make it easier to separate overlapping upper-layer particles.

[0042] Furthermore, the determination unit 16 extracts, as a single separated particle 12, an upper particle of the target portion (or the second target portion) where adjacent particle separation was performed in step S8 (step 9). The determination unit 16 performs circularity determination for each of the target portions where adjacent particle separation was performed using the same shape feature amount (shape coefficient or circularity) as in step S5 (third circularity determination). That is, the determination unit 16 determines which of the target portions where adjacent particle separation was performed is an upper particle based on whether the shape is close to a circle or an ellipse. The determination unit 16 determines, among the target portions where adjacent particle separation was performed, a portion where the calculated shape feature amount is equal to or less than a threshold as an upper particle, and a portion where the calculated shape feature amount is greater than the threshold as a lower particle. The determination unit 16 then extracts the portion determined to be an upper particle as a single separated particle 12. At this time, the determination unit 16 may utilize the threshold used in the circularity determination in steps S5 and S6.

[0043] After the process of step S9, the determination unit 16 proceeds to the process of step S10. Furthermore, if there is no recess in the target portion (No in step S7), the determination unit 16 determines that there is no erroneous detection of stacked particles, and proceeds to the process of step S10.

[0044] If the circularity determination has not been performed for some of the target portions (No in step S10), the determination unit 16 changes the target portions and returns to the processing in step S5.

[0045] When the determination unit 16 has performed the circularity determination for all portions (Yes in step S10), it detects the contours of the identified particles 12. Then, the output unit 17 outputs the contour detection result (step S11, output step). The contour detection result may be, for example, an image showing the contours of the particles 12 (see FIG. 6). The output unit 17 may output the image showing the contours of the particles 12 so that it is displayed on a display connected to a computer functioning as the contour detection device 14. Then, based on the displayed contours of the particles 12, it may be determined whether the particle size of the particles 12 (briquetted coal) is within a reference range. In other words, the contour detection result output by the output unit 17 can be used in an inspection to measure the particle size of the briquetted coal. Here, the processes of steps S3 to S10 executed by the determination unit 16 correspond to the determination step.

[0046] The effects of the present disclosure will be specifically described below based on examples, but the present disclosure is not limited to these examples.

[0047] Fig. 5 shows the results of contour detection in a comparative example in which particles 12 were identified using a conventional contour detection method. Fig. 6 shows the results of contour detection in an example in which particles 12 were identified using the contour detection method according to this embodiment. Figs. 5 and 6 show the results of identifying particles 12 from the same two-dimensional image of conveyed molded coal.

[0048] In the comparative example, binarization was performed on a two-dimensional image in which brightness unevenness was removed by shading correction, and particles 12 were identified using the watershed method. In the example, in addition to the processing of the comparative example, processing for determining circularity and separating stacked particles was performed. That is, in the example, processing by the contour detection method of Figure 3 was performed. As shown in Figure 5, in the comparative example, the stacked particles were not separated, and the contours with recesses can be confirmed. On the other hand, as shown in Figure 6, in the example, the stacked particles were separated, and it can be seen that particles 12 were detected with little error. Here, in Figure 6, for ease of viewing, the contours of the lower-layer particles among the separated stacked particles are not shown.

[0049] As is clear from the comparison with the comparative example, the contour detection method and contour detection system 10 according to this embodiment, with the above-described configuration, can correctly distinguish between upper and lower layer particles in a stack, and can detect particles 12 with little error. The contour detection method and contour detection system 10 according to this embodiment can be used, for example, in an inspection to measure the particle size of molded coal.

[0050] Although the embodiments of the present disclosure have been described based on the drawings and examples, it should be noted that those skilled in the art would easily be able to make various modifications or alterations based on the present disclosure. Therefore, it should be noted that these modifications and alterations are included within the scope of the present disclosure. For example, the functions included in each component or step can be rearranged so as not to cause logical inconsistencies, and multiple components or steps can be combined or divided into one. The embodiments of the present disclosure can also be realized as a program executed by a processor included in an apparatus or a storage medium on which a program is recorded. It should be understood that these are also included within the scope of the present disclosure.

[0051] In the above embodiment, the particles 12 are identified from the brightness of the two-dimensional image output from the CCD camera. As another example, the measurement device 13 may be a laser rangefinder or the like, which performs three-dimensional measurement, and the height information of the three-dimensional data, which is the measurement result, may be used as pixel intensity to identify the particles 12 and detect their contours.

[0052] Furthermore, the contour detection device 14 may not be a single device, but may be composed of multiple devices located in multiple locations and capable of transmitting and receiving data to and from each other via a network. In other words, multiple devices connected via a network may function as a whole as the contour detection device 14. Therefore, for example, the contour detection device 14 may be composed of a single computer as a hardware configuration, or may be composed of multiple computers connected via a network. When composed of multiple computers, a shared memory accessible by each computer may be used to share data or programs.

[0053] Furthermore, when the control device of the measuring device 13 and the contour detection device 14 are configured as a computer, one or more programs used to control the operation of the contour detection system 10 may be stored in a storage device (e.g., memory) of the computer. When the program stored in the storage device is read by a processor included in the computer, the program may cause the processor to function as the control device of the measuring device 13, the acquisition unit 15, the determination unit 16, and the output unit 17. Then, the processing of the contour detection method may be executed by the computer. [Explanation of symbols]

[0054] 10 Contour detection system 11 Conveyor belt 12 particles 12a particles 12b particle 13 Measuring equipment 14 Contour detection device 15 Acquisition Department 16 Judgment section 17 Output section 20 Inscribed Circle 21 Area enclosed by a trajectory line

Claims

1. A contour detection method performed by a contour detection system for detecting contours of stacked circular particles, comprising: A contour detection method comprising: a measurement step of performing two-dimensional or three-dimensional measurement of the particle; an acquisition step of acquiring measurement results of the shape of the particle; and a determination step of identifying the particle based on a shape feature that is an index of a circular shape in the measurement results and pixel intensity that is an index of particle region identification, and detecting the contour of the particle.

2. The contour detection method according to claim 1 , wherein the measurement result is a two-dimensional image or three-dimensional data.

3. 3. The contour detection method according to claim 1, wherein the determining step includes a process of emphasizing boundaries of the particles and a process of separating stacked particles in which a plurality of the particles are stacked.

4. 4. The contour detection method according to claim 3, wherein the determining step performs a first adjacent particle separation on the measurement result after processing to emphasize the boundaries of the particles.

5. The contour detection method of claim 4, wherein the process of separating the stacked particles includes performing a second adjacent particle separation on a target portion that is a candidate for a particle separated by the first adjacent particle separation when the target portion has one or more depressions.

6. 3. The method of claim 1, wherein the pixel intensity is a luminance when the measurement result is a two-dimensional image, and a height when the measurement result is three-dimensional data.

7. 3. The contour detection method according to claim 1, wherein the shape feature amount is a shape coefficient that is an index that objectively indicates whether the shape is close to a circle or an ellipse.

8. The determination step includes: performing a first adjacent particle separation based on the pixel intensities to separate stacked particles of adjacent particles; The contour detection method according to claim 1, further comprising: performing a first circularity determination based on the shape feature to determine whether a first target portion, which is a candidate particle separated by the first adjacent particle separation, is a circular object.

9. The determination step includes: The contour detection method according to claim 8, further comprising: performing a second circularity determination to determine whether the first target portion is a circular object based on whether the first target portion has one or more recesses.

10. 10. The contour detection method according to claim 9, wherein the second circularity determination comprises moving an inscribed circle inscribed in the contour of the first target portion along the contour, and identifying recesses in the first target portion based on the number of areas enclosed by the trajectory of the inscribed circle.

11. The determination step includes: performing a second adjacent particle separation on the first target portion when the second circularity determination determines that the first target portion has one or more recesses; The contour detection method according to claim 9, further comprising a third circularity determination being performed to determine whether a second target portion, which is a candidate particle separated by the second adjacent particle separation, is a circular object based on the shape feature.

12. The determination step includes: The contour detection method according to claim 11 , wherein contours of the first target portion and the second target portion determined to be circular are output as contours of stacked particles.

13. 1. A contour detection system for detecting contours of stacked circular particles, comprising: a measuring device that performs two-dimensional or three-dimensional measurement of the particles; A contour detection system comprising a contour detection device having an acquisition unit that acquires measurement results of the shape of the particle, and a determination unit that identifies the particle based on a shape feature that is an indicator of a circular shape in the measurement results and pixel intensity that is an indicator of particle region identification, and detects the contour of the particle.

Images