Burr detection device, burr removal system, burr detection method, and computer-readable program

Abstract

Regardless of the shape of the object, burrs can be detected accurately and easily. [Solution] The burr detection device 2 for detecting burrs on an object 9 includes a shape acquisition unit 21 that acquires three-dimensional shape information of an inspection area on the object 9, and a burr identification unit 22 that identifies a top area in the inspection area based on the vertical height of each position in the inspection area indicated by the three-dimensional shape information, and identifies a portion as a burr if the cross-sectional area perpendicular to the vertical direction or the width corresponding to the short side of the top area, which is a set length continuous downward from the top area in the vertical direction, satisfies predetermined conditions. The burr detection device 2 can detect burrs accurately and easily, regardless of the shape of the object 9.

Description

Technical Field

[0001] The present invention relates to a technique for detecting burrs on an object.

Background Art

[0002] Conventionally, metal mesh plates have been used for various applications. The mesh plate is cut into a required shape by laser processing or the like. At this time, burrs may occur on the outer edge of the mesh plate formed by cutting. For example, an inspector performs visual and palpation inspections, and repair work is carried out on the locations determined to be burrs. In recent years, inspections by machines have also been carried out. For example, in a contour shape measuring machine, burrs are detected by tracing a stylus on the surface of the target surface to measure the contour shape. There are also increasing numbers of measuring machines that measure the contour non-contact by using a laser instead of a stylus.

[0003] In Patent Document 1, an appearance inspection method is disclosed. First, in a two-dimensional image of the inspected object, a region different from a predetermined luminance value is detected as a defect candidate portion. After removing the defect candidate portion from the three-dimensional data of the inspected object, the three-dimensional data of the region corresponding to the defect candidate portion is interpolated from the data of the peripheral portion. Then, the presence or absence of a defect in the defect candidate portion is determined based on the difference value between the interpolated three-dimensional data and the original three-dimensional data. In Patent Document 2, an appearance inspection device for a soldered state is disclosed. Sequential light irradiation is performed on the soldered portion with illumination having different angles, and each image is captured from obliquely above, and shape information data is created using these images. In the inspection device of Patent Document 3, a three-dimensional planar image showing height information as contour lines is displayed. In Patent Document 4, the operator moves the coordinate axes of the three-dimensional image displayed on the display to change the display range, and the height of the burrs and the like are evaluated by displaying only the burrs on the display.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

[0005] However, visual and tactile inspections for detecting burrs are susceptible to variations in inspection quality depending on the inspector. Even with machine-based inspections, the mesh plates themselves are often warped, requiring knowledge and skills in inspection procedures such as leveling to accurately determine the presence of burrs. Therefore, there is a need for a method that can detect burrs accurately and easily, regardless of the shape of the object.

[0006] This invention has been made in view of the above problems, and aims to accurately and easily detect burrs regardless of the shape of the object. [Means for solving the problem]

[0007] One aspect of the present invention is a burr detection device for detecting burrs on an object, comprising: a shape acquisition unit that acquires three-dimensional shape information of an inspection area on the object; and a burr identification unit that identifies a top area present in the inspection area based on the vertical height of each position in the inspection area indicated by the three-dimensional shape information, and identifies a portion as a burr if the cross-sectional area perpendicular to the vertical direction or the width corresponding to the short side of the top area, for a set length portion continuous downward from the top area in the vertical direction, satisfies predetermined conditions.

[0008] Aspect 2 of the present invention is a burr detection device according to aspect 1, wherein the burr identification unit obtains the vertical length of the burr by using a length in which the cross-sectional area or width satisfies the predetermined conditions, continuously extending downward from the top region in the vertical direction with respect to the portion identified as a burr.

[0009] A third aspect of the present invention is a burr detection device according to aspect 1 (or aspect 1 or 2), wherein the burr identification unit identifies the top region as the cross-sectional region in which the cross-sectional area perpendicular to the vertical direction first falls within a predetermined range when viewed from top to bottom.

[0010] Aspect 4 of the present invention is a burr detection device according to aspect 1 (which may be any one of aspects 1 to 3), wherein the burr identification unit determines either the cross-sectional area or the width as a determination index used to identify the burr, according to the shape of the top region.

[0011] Aspect 5 of the present invention is a burr detection device according to aspect 1 (which may be any one of aspects 1 to 4), wherein the object is a plate-shaped member or a mesh plate having deflection in an unspecified direction.

[0012] Aspect 6 of the present invention is a burr removal system comprising a burr detection device according to any one of aspects 1 to 5, and a burr removal device for removing burrs detected by the burr detection device.

[0013] Aspect 7 of the present invention is a burr detection method for detecting burrs on an object, comprising the steps of: acquiring three-dimensional shape information of an inspection area on the object; identifying a top area in the inspection area based on the vertical height of each position in the inspection area indicated by the three-dimensional shape information; and identifying a portion of a set length that is continuous downward from the top area in the vertical direction as a burr if the cross-sectional area perpendicular to the vertical direction or the width corresponding to the short side of the top area satisfies predetermined conditions.

[0014] Aspect 8 of the present invention is a computer-readable program that causes a computer to detect burrs on an object. The execution of the program by the computer causes the computer to perform the steps of: preparing three-dimensional shape information of an inspection area on the object; identifying a top area existing in the inspection area based on the vertical height of each position in the inspection area indicated by the three-dimensional shape information; and identifying a portion having a set length that is continuous downward in the vertical direction from the top area as a burr when the cross-sectional area perpendicular to the vertical direction or the width corresponding to the short side direction of the top area satisfies a predetermined condition.

Advantages of the Invention

[0015] According to the present invention, burrs can be accurately and easily detected regardless of the shape of the object.

Brief Description of the Drawings

[0016] [Figure 1] It is a diagram showing the configuration of a deburring system. [Figure 2] It is a plan view of a mesh plate. [Figure 3] It is a side view of a mesh plate. [Figure 4] It is a diagram showing the configuration of a computer. [Figure 5] It is a diagram showing the flow of a process for detecting and removing burrs on a mesh plate. [Figure 6] It is a diagram showing a part of an image indicated by three-dimensional shape information. [Figure 7] It is a diagram showing a contour image. [Figure 8A] It is a diagram for explaining the identification of burrs by a burr identification unit. [Figure 8B] It is a diagram for explaining the identification of burrs by a burr identification unit. [Figure 9] It is a diagram showing another example of an object. [Figure 10] It is a diagram showing another example of a deburring system. [Figure 11] It is a diagram showing another example of a deburring system. [Modes for carrying out the invention]

[0017] Figure 1 shows the configuration of a burr removal system 1 according to one embodiment of the present invention. The burr removal system 1 detects burrs on an object and removes them. In this embodiment, the object is a mesh plate 9. In the example in Figure 1, the mesh plate 9 is placed on a substantially horizontal surface (belt 51, described later). Hereinafter, for the sake of convenience in explanation, the direction substantially perpendicular to the mesh plate 9 will be referred to as the "up and down direction," but the up and down direction is not limited to the direction of gravity.

[0018] Figure 2 is a plan view of the mesh plate 9, showing the mesh plate 9 as viewed along the vertical direction. The mesh plate 9 is a mesh-like plate made of metal or the like, and extends approximately vertically. In the example in Figure 2, the outer shape of the mesh plate 9 is approximately rhombic, but it may have other shapes. The mesh plate 9 may be, for example, expanded metal, perforated metal, woven mesh, etc., and has numerous holes.

[0019] Figure 3 is a side view of the mesh plate 9, showing it as viewed from a direction perpendicular to one edge of the outer edge of the mesh plate 9 and also perpendicular to the vertical direction. The burrs 91 are unwanted protrusions present on the outer edge of the mesh plate 9, and are typically generated during the cutting process of the outer edge of the mesh plate 9. The burrs 91 generated on the mesh plate 9 are needle-shaped. Some of the burrs 91 protrude upward, while others protrude downward. In the following description, the part of the mesh plate 9 excluding the burrs 91 will also be referred to as the "main body". As shown in Figure 1, the main body of many mesh plates 9 has deflection in an unspecified direction (here, deformation when no load is applied). The state (direction and magnitude) of deflection in the mesh plate 9 varies.

[0020] The deburring system 1 shown in Figure 1 comprises a deburring detection device 2, a deburring device 4, and a conveying unit 5. The conveying unit 5 is, for example, a belt conveyor, and moves the mesh plate 9 placed on the belt 51 by circulating the belt 51. In the example in Figure 1, the surface of the belt 51 on which the mesh plate 9 is placed is substantially perpendicular in the vertical direction, and the conveying unit 5 moves the mesh plate 9 in the horizontal direction of Figure 1. The conveying unit 5 may be something other than a belt conveyor, such as a roller conveyor.

[0021] The burr detection device 2 comprises a shape acquisition unit 21, a robot 11, and a computer 3. The robot 11 is, for example, a vertical articulated robot, and in the example shown in Figure 1, it comprises an arm 111 having 6 joints (6 axes). The number of joints in the arm 111 may be 5 or less, or 7 or more. The robot 11 may also be something other than a vertical articulated robot. Since the robot 11 is a well-known industrial robot, a detailed explanation is omitted.

[0022] The shape acquisition unit 21 is, for example, a digital microscope and acquires three-dimensional shape information of the mesh plate 9. The shape acquisition unit 21 is attached to the end of the arm 111 of the robot 11. When measurement is performed by the shape acquisition unit 21, the robot 11 positions the shape acquisition unit 21 opposite the inspection area on the mesh plate 9. At this time, the direction of the optical axis of the shape acquisition unit 21 is approximately perpendicular to the surface of the mesh plate 9, and by moving the lens or the like in the optical axis direction, the height at which the focus is achieved for each position in the field of view is determined. This acquires three-dimensional shape information of the inspection area on the mesh plate 9. The shape acquisition unit 21 may be something other than a digital microscope, for example, a laser scanner. Typically, the shape acquisition unit 21 acquires three-dimensional shape information of the mesh plate 9 without contact.

[0023] Figure 4 shows the configuration of computer 3. Computer 3 has the configuration of a general computer system, including a CPU 31 that performs various calculations, a ROM 32 that stores basic programs, and a RAM 33 that stores various information. Computer 3 further includes a GPU 39 that performs various calculations related to image processing, a storage device 34 that stores information, a display unit 35 that displays various information such as images, a keyboard 36a and a mouse 36b (hereinafter collectively referred to as "input unit 36") that accept input from the operator, a reader 37 that reads information from computer-readable recording media 8 such as optical discs, magnetic discs, magneto-optical discs, and memory cards, and a communication unit 38 that sends and receives signals with other components of the deburring system 1.

[0024] In computer 3, program 80 is read in advance from recording medium 8 via reading device 37 and stored in storage device 34. Program 80 may also be stored in storage device 34 via a network. CPU 31 and GPU 39 perform calculations using RAM 33 and storage device 34 according to program 80. This realizes the burr identification unit 22 in Figure 1. All or part of the burr identification unit 22 may be realized by a dedicated electrical circuit. Alternatively, the burr identification unit 22 may be realized by multiple computers. The burr identification unit 22 identifies burrs 91 on the mesh plate 9 based on three-dimensional shape information acquired by shape acquisition unit 21. Computer 3 is also responsible for the overall control of the burr removal system 1.

[0025] The deburring device 4 comprises a deburring unit 41 and a robot 11. The robot 11 is shared by the deburring detection device 2 and the deburring device 4. The deburring unit 41 is, for example, a deburring spindle that rotates a grinding wheel or wire brush provided at its tip. The deburring unit 41 is attached to the end of the arm 111 of the robot 11. The robot 11 removes the burrs 91 from the mesh plate 9 by bringing the tip of the deburring unit 41 into contact with the burrs 91 on the mesh plate 9. The deburring unit 41 may also be a cutting unit that cuts the burrs 91 from the base.

[0026] Figure 5 shows the flow of the process by which the deburring system 1 detects and removes burrs 91 from the mesh plate 9. First, the transport unit 5 in Figure 1 drives the belt 51, so that one mesh plate 9 is placed at a predetermined inspection position on the transport path. The inspection position is opposite the shape acquisition unit 21 in Figure 1, and the mesh plate 9 stops at the inspection position.

[0027] Next, the shape acquisition unit 21 acquires three-dimensional shape information for the inspection area of ​​the mesh plate 9 (step S11). In this example, the inspection area is the entire circumference of the outer edge of the mesh plate 9, and in Figure 2, parallel diagonal lines are drawn over the inspection area R1. Above the mesh plate 9, the robot 11 moves the shape acquisition unit 21 along the outer edge of the mesh plate 9 (see arrow A1 in Figure 2), thereby acquiring three-dimensional shape information for the entire inspection area R1. In this embodiment, the shape acquisition unit 21 acquires a grayscale image as three-dimensional shape information, which shows the vertical height of each position in the inspection area R1 as gradation values. For example, one pixel in the image is 1 to 100 μm. 2 This corresponds to a vertical distance of 1 to 100 μm, where the smallest unit of the grayscale value corresponds to a vertical distance of 1 to 100 μm. Of course, the size corresponding to one pixel and the distance corresponding to the smallest unit of the grayscale value can be determined arbitrarily. The 3D shape information is input to the computer 3 and prepared for processing in the burr identification unit 22.

[0028] Figure 6 shows an image represented by 3D shape information, and shows a part of the inspection area R1 of the mesh plate 9. In the image in Figure 6, the higher the vertical height of each position in the inspection area R1 (located higher up), the larger the grayscale value of the pixel at that position, and the brighter the pixel becomes. The lower the vertical height of each position (located lower down), the smaller the grayscale value of the pixel at that position, and the darker the pixel becomes. In the burr identification unit 22, for each grayscale value in the image represented by 3D shape information, lines (i.e., contour lines) are set to surround groups of pixels with a grayscale value equal to or greater than that value, thereby obtaining an image that shows the height information of the inspection area R1 using contour lines (hereinafter referred to as the "contour image"), as shown in Figure 7. In the following explanation, the area enclosed by each contour line will simply be referred to as the "contour area".

[0029] Next, the burr identification unit 22 identifies the top region present in the inspection area R1 (step S12). The top region is the part that is most likely to be the tip of the burr 91. Figure 8A is a diagram illustrating the identification of the burr 91 by the burr identification unit 22, and shows the cross-sectional shapes C1 to C4 at four different positions on the mesh plate 9 side by side. In Figure 8A, the vertical axis represents the height in the vertical direction, and the horizontal axis represents the direction perpendicular to the vertical direction. In addition, the height corresponding to each grayscale value is shown by a thin dashed line parallel to the horizontal axis.

[0030] In this processing example using contour images, the number of pixels in the region of the locally highest contour line (i.e., the vertex) is compared with the vertex determination range. The vertex determination range is, for example, 100 to 1000. The vertex determination range can be determined arbitrarily. If the number of pixels in the region falls within the vertex determination range, the region is identified as a vertex region. If the number of pixels in the region is greater than the vertex determination range, the vertex region is not identified. If the number of pixels in the region is less than the vertex determination range, the number of pixels in the region of the next highest contour line (i.e., the contour line adjacent to the lower side) is compared with the vertex determination range. If the number of pixels in the region falls within the vertex determination range, the region is identified as a vertex region. If the number of pixels in the region is greater than the vertex determination range, the vertex region is not identified. If the number of pixels in the region is less than the vertex determination range, the number of pixels in the region of the contour line adjacent to the lower side of the contour line is compared with the vertex determination range in the same manner as above.

[0031] As described above, in the burr identification unit 22, the cross-sectional area that is vertically perpendicular to the top and bottom, and is the first to fall within a predetermined range (top determination range) from top to bottom at height intervals corresponding to the smallest unit of grayscale value, is identified as the top region. In the example in Figure 8A, in the parts of the mesh plate 9 represented by the leftmost cross-sectional shape C1 and the third cross-sectional shape C3 from the left, the number of pixels in the locally highest contour area is greater than the top determination range, so the top region is not identified. Such parts are considered to be the main body of the mesh plate 9, not burrs. In the parts of the mesh plate 9 represented by the second cross-sectional shape C2 from the left and the rightmost cross-sectional shape C4, the number of pixels in the locally highest contour area is relatively small and falls within the top determination range, so it is identified as the top region. In Figure 8A, the range of the top region is indicated by an arrow with the reference numeral 92, and in the following explanation as well, the top region will be indicated by the reference numeral 92.

[0032] Figure 8B is a diagram illustrating the identification of burrs 91 by the burr identification unit 22, and shows only the cross-sectional shapes C2 and C4 in Figure 8A. In Figure 8B, dashed lines indicating the height corresponding to each grayscale value are drawn only inside the cross-sectional shapes C2 and C4. The length of the dashed lines inside the cross-sectional shapes C2 and C4 corresponds to the width of the contour line region in the cross-section from which the cross-sectional shapes C2 and C4 are obtained. In the burr identification unit 22, for each top region 92, the number of pixels in the region of the set number of contour lines that are continuous below the contour line of the top region 92 is compared with the set number of pixels. For example, the set number of contour lines is 10, and the set number of pixels is 2000. Of course, the set number of contour lines and the set number of pixels may be other values. In Figure 8B, the vertical range including the set number of contour lines from the contour line of the top region 92 is indicated by arrow E1. If the number of pixels is less than or equal to the set number of pixels in all areas of the contour lines of the set number, the top area 92 and the portion including the contour lines of the set number are identified as a variance 91 (step S13). For example, in the cross-sectional shape C2 in Figure 8B, the width of the contour line area is relatively small throughout the entire range E1, and the number of pixels is less than or equal to the set number of pixels in all areas of the contour lines of the set number. Therefore, the portion of cross-sectional shape C2 is identified as a variance 91. On the other hand, if the number of pixels is greater than the set number of pixels in any of the contour line areas of the set number, the top area 92 and the portion including the contour line area are determined not to be a variance 91. For example, in the upper part of range E1 in the cross-sectional shape C4 in Figure 8B, the width of the contour line area is relatively small, and the number of pixels in the contour line area is less than or equal to the set number of pixels. On the other hand, in the lower part of range E1, which is the main body of the mesh plate 9, the width of the contour line area is very large, and the number of pixels in the contour line area far exceeds the set number of pixels. Thus, in cross-sectional shape C4, the number of contour lines where the number of pixels is less than or equal to the set number is less than the set number, so the portion of cross-sectional shape C4 is determined not to be a 91 contour line.

[0033] If we refer to the vertical length corresponding to the set number of contour lines as the "set length," then step S13 can be described as a process that identifies a portion of the set length that extends vertically downward from the top region 92 as a variance 91 if the vertically perpendicular cross-sectional area of ​​that portion is within a set value (a value corresponding to the set number of pixels). In step S13, the portion of the set length may also be identified as a variance 91 if the ratio of the cross-sectional area of ​​each contour line region included in the set length to the cross-sectional area of ​​the top region 92 is within a threshold, or if the rate of change of the cross-sectional area of ​​each contour line region (the rate of change from the cross-sectional area of ​​the contour line region adjacent to the upper side) is within a threshold, etc. Thus, the variance identification unit 22 identifies a portion of the set length that extends vertically downward from the top region 92 as a variance 91 if the vertically perpendicular cross-sectional area satisfies a predetermined condition (typically, a condition that it does not become excessively large).

[0034] In the preferred burr detection device 2, when a burr 91 is identified, the number of contour lines that satisfy the condition (in this case, the number of pixels is less than or equal to the set number of pixels) that are continuous below the contour lines of the top region 92 is also determined. The value obtained by multiplying this number by the height interval corresponding to the smallest unit of grayscale value is acquired as the vertical length of the burr 91. In the second cross-sectional shape C2 from the left in Figure 8A, the vertical length of the burr 91 is indicated by arrow L1 (similarly in the cross-sectional shape C2 in Figure 8B). The vertical length of the burr 91 may also be determined by adding the length of the part above the top region 92.

[0035] When a burr 91 is identified, the robot 11 brings the burr removal unit 41 into contact with the burr 91, thereby removing the burr 91 (step S14). Once all burrs 91 identified by the burr identification unit 22 have been removed by the burr removal unit 41, the processing of the mesh plate 9 placed at the inspection position is completed. Subsequently, the transport unit 5 drives the belt 51 to place the next mesh plate 9 at the inspection position, and steps S11 to S14 are repeated for that mesh plate 9.

[0036] In steps S11 to S14 described above, only the burrs 91 protruding upward in Figure 3 are removed, while the burrs 91 protruding downward are not removed. In the burr removal system 1, the mesh plate 9 that has undergone the processing in steps S11 to S14 is inverted and placed back onto the belt 51, and the processing in steps S11 to S14 is repeated. This removes the remaining burrs 91 (the burrs 91 that protruded downward in the initial processing).

[0037] Table 1 shows an example of the measurement results of the vertical length of burrs 91 obtained by the burr detection device 2. In this measurement, as in the processing example above, the number of contour lines where the number of pixels is less than or equal to the set number of pixels (here, 2000) was determined, and the value obtained by multiplying this number by the height interval corresponding to the smallest unit of grayscale value was obtained as the estimated burr length. In addition, each burr 91 for which the estimated burr length was obtained was imaged from a direction perpendicular to the vertical direction, and the vertical length of the burr 91 measured in the captured image was determined as the true burr length. As shown in Table 1, the difference between the estimated burr length and the actual burr length was small for the three burrs measured in this test, indicating that the burr detection device 2 can accurately measure the vertical length of burrs 91.

[0038] [Table 1]

[0039] As described above, the burr detection device 2 comprises a shape acquisition unit 21 and a burr identification unit 22. The shape acquisition unit 21 acquires three-dimensional shape information of the inspection area R1 on the object (mesh plate 9 in the above case). The burr identification unit 22 identifies the top area 92 present in the inspection area R1 based on the vertical height of each position in the inspection area R1 indicated by the three-dimensional shape information. Furthermore, for a portion of a set length that is continuous downward from the top area 92 in the vertical direction, if the cross-sectional area perpendicular to the vertical direction satisfies predetermined conditions, that portion is identified as a burr 91.

[0040] Here, we consider the burr detection device of the comparative example. In the burr detection device of the comparative example, the surface on which the object is placed is used as the reference plane, and the height from the reference plane is measured for each part of the object. When the object is a plate-shaped member that deflects in an unspecified direction, such as the mesh plate 9, the measured height of each part includes the effect of deflection (i.e., the height of the main body of the object itself is not constant), making it difficult to detect burrs from these measured values ​​alone. Furthermore, Japanese Patent Publication No. 4894628 (Patent Document 1 above) discloses a method in which, after removing candidate defect parts from the 3D data of the object, the 3D data of that region is interpolated from the data of the surrounding parts, and the difference value from the original 3D data is obtained, and it is conceivable that this method could be applied to burr detection. However, with this method, when the object has discontinuous parts and a complex shape, such as the mesh plate 9, the amount of computation becomes enormous, or interpolation becomes difficult.

[0041] In contrast, the burr detection device 2 determines whether a part is a burr 91 by examining the cross-sectional area of ​​the part where the top region 92 has been identified, moving downward from the top region 92. As a result, it becomes possible to detect burrs 91 accurately and easily, regardless of the shape of the object (even if the object has a complex shape).

[0042] Preferably, the burr identification unit 22 obtains the vertical length of the burr 91 by using a length that satisfies a predetermined condition for the cross-sectional area, continuously extending downward from the top region 92 in the vertical direction for the portion identified as the burr 91. This allows the vertical length of the burr 91 to be obtained with high accuracy.

[0043] The deburring system 1 comprises a deburring detection device 2 and a deburring device 4 that removes the deburring detected by the deburring detection device 2. This allows for the precise removal of deburring 91 from the object.

[0044] Incidentally, in the processing of step S12 in the burr identification unit 22, it is also possible to identify all areas of the locally highest contour lines as peak regions. Even in this case, in the processing of step S13, it is possible to appropriately detect the burr 91 by checking whether the cross-sectional area of ​​a set length portion that is continuous downward in the vertical direction from each peak region satisfies predetermined conditions. On the other hand, as in the processing example above, if the burr identification unit 22 identifies the cross-sectional region where the cross-sectional area perpendicular to the vertical direction first falls within a predetermined range from top to bottom as the peak region 92, it is possible to exclude peak regions that are clearly not the tip of the burr 91 from the processing of step S13. In this way, by appropriately identifying the peak region 92, it is possible to reduce the amount of calculation in the burr identification unit 22.

[0045] Figure 9 shows another example of the object, and shows a portion of a metal plate 9a. The metal plate 9a in Figure 9 extends substantially vertically in the vertical direction and has, for example, a rectangular outline. The metal plate 9a may have some curvature. The metal plate 9a has a burr 91 that extends along the outer edge and protrudes upward in Figure 9. In the example in Figure 9, the burr 91 is plate-shaped.

[0046] When the burr removal system 1 detects and removes burrs 91 from the metal plate 9a, the shape acquisition unit 21 acquires three-dimensional shape information for the inspection area of ​​the metal plate 9a (Figure 5: Step S11). This inspection area is, for example, the entire circumference of the outer edge of the metal plate 9a. In the burr identification unit 22, similar to the processing example above, if the number of pixels in a region with locally high contour lines in the contour image derived from the three-dimensional shape information falls within the peak determination range, that region is identified as the peak region 92 (Step S12). In Figure 9, the peak region 92 is shown enclosed by a dashed line.

[0047] Once each vertex region 92 is identified, the short-side direction of that vertex region 92 is determined. For example, a circumscribing rectangle is set for the vertex region 92, and the direction parallel to the short side of the circumscribing rectangle is determined as the short-side direction. Typically, the short-side direction is perpendicular to the vertical direction and approximately perpendicular to the outer edge of the metal plate 9a. The short-side direction may be determined by other methods; for example, an elliptic approximation may be performed on the vertex region 92, and the direction of the minor axis may be determined as the short-side direction.

[0048] Next, in the region of a set number of contour lines that are continuous downward from the contour lines of the top region 92, the width corresponding to the shorter direction of the top region 92 (the width in that shorter direction, hereinafter referred to as "shorter width") is compared with a set value. The set value is, for example, the upper limit of the shorter width. In Figure 9, for reference, the width in the shorter direction of the top region 92 is shown by arrow W1. If the shorter width is within the set value in all of the regions of the set number of contour lines, the top region 92 and the portion including the region of the set number of contour lines (i.e., the portion of the set length continuous downward from the top region 92) are identified as a variance 91 (step S13). On the other hand, if the shorter width is greater than the set value in any of the regions of the set number of contour lines, the portion of the set length continuous from the top region 92 is determined not to be a variance 91.

[0049] In step S13, the portion of the set length may be identified as a burr 91 if the ratio of the short-width of each contour region included in the set length to the short-width of the top region 92 is within a threshold, or if the rate of change of the short-width of each contour region (the rate of change from the short-width of the contour region adjacent to the upper side) is within a threshold. In this way, the burr identification unit 22 identifies the portion of the set length that is continuous downward in the vertical direction from the top region 92 as a burr 91 if the short-width satisfies a predetermined condition (typically, a condition that it does not become excessively large).

[0050] In the preferred burr detection device 2, when a burr 91 is identified, the vertical length of the burr 91 is also obtained by determining the number of contour lines that satisfy the condition (in this case, the short-width is within a set value) that are continuous below the contour line of the top region 92. As previously described, the vertical length of the burr 91 may also be determined by adding the length of the part above the top region 92. The burr 91 identified by the burr identification unit 22 is removed by the burr removal unit 41 (step S14).

[0051] As described above, the burr identification unit 22 identifies a portion of a set length that extends downward from the top region 92 in the vertical direction, as a burr 91, if the width corresponding to the shorter side of the top region 92 (short side width) satisfies a predetermined condition. In this way, by examining the short side width of the portion of the identified top region 92 in the downward direction from the top region 92, burrs 91 extending along the outer edge of the metal plate 9a can be detected accurately and easily. Furthermore, for the portion identified as a burr 91, the vertical length of the burr 91 is obtained using a length that extends downward from the top region 92 in the vertical direction and satisfies the predetermined condition for the short side width. This allows for the accurate acquisition of the vertical length of the burr 91.

[0052] The outer shape of the metal plate 9a may not be rectangular. For example, if the outer shape of the metal plate 9a includes an arc-shaped portion, the top region 92 is approximated by the arc region, and the radial direction of the arc region is determined as the short side direction. Then, for a portion of a set length that is continuous below the top region 92 in the vertical direction, if the width corresponding to the short side direction of the top region 92 (the radial width when the cross section perpendicular to the vertical direction is approximated by the arc region) satisfies a predetermined condition, that portion is identified as a burr 91.

[0053] In the above processing example, for a mesh plate 9 where needle-shaped burrs 91 occur as shown in Figure 3, the burrs 91 are identified using the cross-sectional area perpendicular to the vertical direction as a determination index, and for a metal plate 9a where plate-shaped burrs 91 occur as shown in Figure 9, the burrs 91 are identified using the width of the shorter side as a determination index. For example, if needle-shaped burrs 91 and plate-shaped burrs 91 are mixed in the object, the burr identification unit 22 may determine either the cross-sectional area or the width of the shorter side as a determination index to be used to identify the burrs 91, according to the shape of each top region 92.

[0054] For example, a circumscribing rectangle is defined around the top region 92, and the ratio of (length of the long side) / (length of the short side) in the circumscribing rectangle is determined. If the value of this ratio is greater than or equal to a predetermined value, the width of the short side is used as the judgment index for the top region 92. If the value of this ratio is less than the predetermined value, the cross-sectional area perpendicular to the vertical direction is used as the judgment index for the top region 92. In this way, by using an appropriate judgment index according to the shape of the top region 92, it becomes possible to accurately detect burrs 91 in objects where needle-shaped burrs 91 and plate-shaped burrs 91 are mixed. In determining the judgment index, values ​​other than the lengths of the long and short sides of the circumscribing rectangle may be used. For example, the top region 92 may be approximated as an ellipse, and the ratio of (length of the major axis) / (length of the minor axis) may be determined. Also, if the top region 92 is approximated as a circular arc region, the ratio of (length in the circumferential direction) / (length in the radial direction) may be determined.

[0055] The above-described burr removal system 1, burr detection device 2, and burr detection method can be modified in various ways.

[0056] In the deburring system 1 shown in Figure 1, the shape acquisition unit 21 and the deburring unit 41 are attached to the arm 111 of the robot 11. However, as shown in Figure 10, the shape acquisition unit 21 and the deburring unit 41 may be attached to a moving body of a moving mechanism 11a that can move the moving body in three mutually orthogonal directions. Alternatively, the shape acquisition unit 21 and the deburring unit 41 may be movable independently. In the example shown in Figure 11, the shape acquisition unit 21 is movable by the moving mechanism 11a, and the deburring unit 41 is attached to the arm 111 of the robot 11. In the deburring system 1 shown in Figures 1 and 10, it is possible to remove the burr 91 immediately after detection. Furthermore, since the relative positions of the shape acquisition unit 21 and the deburring unit 41 are fixed, there is no need for operations such as alignment between the two. In the deburring system 1 shown in Figure 11, it is possible to acquire 3D shape information by the shape acquisition unit 21 and remove the burr 91 by the deburring unit 41 in parallel. Furthermore, it eliminates the constraints that arise when both are attached to the same moving mechanism.

[0057] In the above embodiment, burrs 91 formed on the outer edge of the object are detected. However, in cases where holes are formed, such as in perforated metal, burrs in the holes may also be detected by the burr detection device 2. In this case, for example, the top region is approximated by an arc region, and the radial direction of the arc region is used as the short side direction.

[0058] The objects targeted by the burr removal system 1 and the burr detection device 2 are not limited to plate-shaped objects, but may also be three-dimensional in shape, such as machine parts, and may be made of materials other than metal (for example, resin).

[0059] The configurations in the above embodiments and each modified example may be combined as appropriate, as long as they do not contradict each other. [Explanation of Symbols]

[0060] 1. Deburring System 2. Burr detection device 3 Computers 4. Deburring device 9 Mesh board 9a metal plate 21 Shape acquisition section 22 Bali Special Section 80 Programs 91 Bali 92 Top area R1 Examination Area S11~S14 Step

Claims

1. A burr detection device for detecting burrs on an object, A shape acquisition unit that acquires three-dimensional shape information of the inspection area on the object, A burr identification unit identifies a top region in the inspection region based on the vertical height of each position in the inspection region indicated by the three-dimensional shape information, and identifies a portion of a set length that is continuous downward from the top region in the vertical direction as a burr if the cross-sectional area perpendicular to the vertical direction or the width corresponding to the shorter side of the top region satisfies predetermined conditions. A burr detection device equipped with the following features.

2. A burr detection device according to claim 1, A burr detection device that, with respect to the portion identified as a burr by the burr identification unit, obtains the vertical length of the burr using a length in which the cross-sectional area or width satisfies the predetermined conditions, continuously extending downward from the top region in the vertical direction.

3. A burr detection device according to claim 1, A burr detection device in which the burr identification unit identifies the top region as the cross-sectional region in which the cross-sectional area perpendicular to the vertical direction first falls within a predetermined range when viewed from top to bottom.

4. A burr detection device according to claim 1, A burr detection device in which the burr identification unit determines either the cross-sectional area or the width as a determination index used to identify the burr, according to the shape of the top region.

5. A burr detection device according to claim 1, A burr detection device in which the object is a plate-shaped member or mesh plate having deflection in an unspecified direction.

6. It is a deburring system, A burr detection device according to any one of claims 1 to 5, A burr removal device for removing burrs detected by the burr detection device, A deburring system equipped with the following features.

7. A burr detection method for detecting burrs on an object, A process to acquire 3D shape information of the inspection area on the object, A step of identifying the top region present in the inspection region based on the vertical height of each position in the inspection region indicated by the three-dimensional shape information, With respect to a portion of a set length that is continuous downward in the vertical direction from the top region, if the cross-sectional area perpendicular to the vertical direction, or the width corresponding to the shorter side of the top region, satisfies predetermined conditions, the process involves identifying the portion as a burr. A burr detection method comprising the following features.

8. A computer-readable program that causes a computer to perform the detection of burrs on an object, wherein the execution of the program by the computer is performed by the computer, A process to prepare 3D shape information of the inspection area on the object, A step of identifying the top region present in the inspection region based on the vertical height of each position in the inspection region indicated by the three-dimensional shape information, With respect to a portion of a set length that is continuous downward in the vertical direction from the top region, if the cross-sectional area perpendicular to the vertical direction, or the width corresponding to the shorter side of the top region, satisfies predetermined conditions, the process involves identifying the portion as a burr. A computer-readable program that executes.

Images