Image inspection apparatus

Abstract

The image inspection apparatus includes: an image acquisition unit that acquires a captured image obtained by capturing an inspection region by a camera; a camera information storage unit that stores camera distance information between the camera and the inspection region; a drawing information storage unit that stores drawing information including size information of a concave-convex processed surface of the inspection region; a planar development processing unit that converts the captured image of the concave-convex processed surface of the inspection region into a planar development captured image using the camera distance information and the drawing information; an image scale correction unit that calculates an image scale correction value for each concave-convex processed surface of the inspection region using the camera distance information and the drawing information, and generates a scale-corrected image of the planar development captured image using the image scale correction value; and a defect extraction processing unit that extracts a defect from the scale-corrected image.

Description

Technical Field

[0001] This invention relates to an image inspection apparatus. Background Technology

[0002] Patent document 1 discloses an inner circumference inspection device for a cylindrical container, which includes an illumination unit, an imaging unit, and an image processing unit, and takes pictures of the inside of the cylinder while rotating the cylindrical container.

[0003] Existing technical documents

[0004] Patent documents

[0005] Patent Document 1: Japanese Patent Application Publication No. 2012-26858 Summary of the Invention

[0006] The technical problem that the invention aims to solve

[0007] In Patent Document 1, the scales of the cylinder on the front and back sides of the cylinder are different in the captured image, which may make it impossible to extract defects of a specific size.

[0008] One of the objectives of this invention is to provide an image inspection device that can prevent defects displayed due to different scales from being missed, thereby improving the accuracy of defect extraction.

[0009] Technical solutions for solving technical problems

[0010] An image inspection apparatus according to one embodiment of the present invention includes: an image acquisition unit that acquires a captured image of an inspection area by a camera; a camera information storage unit that stores camera distance information between the camera and the inspection area; a drawing information storage unit that stores drawing information having dimension information of the uneven surfaces of the inspection area; a planar unfolding processing unit that uses the camera distance information and the drawing information to convert the captured image of the uneven surfaces of the inspection area into a planar unfolded image; an image scale correction unit that uses the camera distance information and the drawing information to calculate an image scale correction value for each uneven surface of the inspection area, and uses the image scale correction value to generate a scale-corrected image of the planar unfolded image; and a defect extraction processing unit that extracts defects from the scale-corrected image.

[0011] According to one embodiment of the present invention, defects displayed due to different scales can be prevented from being missed, thereby improving the accuracy of defect extraction. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of the overall image inspection device 1 according to Embodiment 1.

[0013] Figure 2This is the main flowchart illustrating the image inspection process in Implementation 1.

[0014] Figure 3 This is a cross-sectional view of the cylindrical hole in Embodiment 1.

[0015] Figure 4 This is a timing diagram showing the operation flow of the planar unfolding processing unit and the image scale correction unit in Embodiment 1.

[0016] Figure 5 This is a timing diagram illustrating the generation process of the virtual boundary line image in Implementation 1.

[0017] Figure 6 This is a timing diagram illustrating the process of obtaining the boundary line of a planar unfolded image of a rectangular image unfolded in the circumferential direction in Embodiment 1.

[0018] Figure 7 (a) and (b) are diagrams illustrating the method for determining that an image capture has been obstructed in Implementation 1.

[0019] Figure 8 This is a flowchart illustrating the process of determining occlusion (masking) and calculating image scaling values ​​in Implementation 1.

[0020] Figure 9 This is a schematic diagram of the overall image inspection device according to Embodiment 2.

[0021] Figure 10 This is the main flowchart illustrating the image inspection process in Implementation Method 2. Detailed Implementation

[0022] [Implementation Method 1]

[0023] Figure 1 This is a schematic diagram of the overall image inspection device 1 according to Embodiment 1.

[0024] The image inspection device 1 of Embodiment 1 includes a aperture lens camera (camera) 2 and a computer 3.

[0025] The aperture lens camera 2 takes pictures of the cylindrical hole portion (inspection area) 17a of the VTC cover 17, which is the object to be inspected, with concave and convex machined surfaces.

[0026] Computer 3, for example, is a personal computer, equipped with CPU 4 and memory 5.

[0027] CPU4 includes a camera communication unit 6, an image acquisition unit 7, a planar unfolding processing unit 8, an image scale correction unit 9, and a defect extraction processing unit 10.

[0028] The camera communication unit 6 controls the camera 2 to capture an image by taking a single shot from the concave and convex machined surface inside the cylindrical hole 17a, and the image acquisition unit 7 acquires the image.

[0029] In this way, only one image needs to be taken of the entire concave and convex machined surface inside the cylindrical hole 17a, thus shortening the shooting time.

[0030] The details of the processing of the planar unfolding processing unit 8, the image scale correction unit 9, and the defect extraction processing unit 10 will be described later.

[0031] The memory 5 includes an image storage unit 11, a camera information storage unit 12, a map information storage unit 13, an image scale correction value storage unit 14, a scale correction image storage unit 15, and a defect extraction result storage unit 16.

[0032] The image storage unit 11 stores the image captured by the image acquisition unit 7 inside the cylindrical hole 17a; the camera information storage unit 12 stores the distance between the aperture lens camera 2 and the cylindrical hole 17a, the resolution of the aperture lens camera 2, etc.; the drawing information storage unit 13 stores drawing information containing the size information (design dimensions, shape data, etc.) of each concave and convex machined surface inside the cylindrical hole 17a; the image scale correction value storage unit 14 stores the image scale correction value; the scale correction image storage unit 15 stores the scale correction image generated by the image scale correction unit 9; and the defect extraction result storage unit 16 stores the defect extraction result of the defect extraction processing unit 10.

[0033] Figure 2 This is the main flowchart illustrating the image inspection process in Implementation 1.

[0034] In step S1, based on the camera distance information stored in the camera information storage unit 12, the image information stored in the image information storage unit 13, and the image obtained by capturing a single shot of the concave and convex machined surface in the cylindrical hole 17a stored in the captured image storage unit 11, the concave and convex machined surface in the cylindrical hole 17a is converted into a planar unfolded captured image.

[0035] In step S2, since the image is taken as a single image of the entire concave and convex machined surface inside the cylindrical hole 17a, the scale is uneven on the near front and inner sides of the cylindrical hole 17a (the scale becomes smaller towards the inner side). Therefore, based on the camera distance information, the image information, and the planar unfolded image of the concave and convex machined surface inside the cylindrical hole 17a, the image scale correction value is calculated according to each concave and convex machined surface inside the cylindrical hole 17a, and a scale-corrected image of the planar unfolded image of the concave and convex machined surface inside the cylindrical hole 17a with a uniform scale is generated.

[0036] In step S3, defects are extracted from the scale-corrected image.

[0037] Figure 3 This is a cross-sectional view of the cylindrical hole in Embodiment 1.

[0038] The concave and convex machined surfaces of the cylindrical hole 17a are composed of a chamfered machined surface 17a1 in the cylindrical direction, an upper cylindrical machined surface 17a2, a grooved machined surface 17a3, upper and lower grooved machined side surfaces 17a4, and a lower cylindrical machined surface 17a5.

[0039] Figure 4 This is a timing diagram showing the processing flow of the planar unfolding processing unit and the image scale correction unit in Embodiment 1.

[0040] In the first block P1, the planar unfolding processing unit 8 obtains the captured image obtained by a single capture and stored in the captured image storage unit 11.

[0041] It should be noted that the grid pattern K (calibration image) of the captured image is projected using lasers or other methods.

[0042] In the second block P2, the captured image obtained by the planar unfolding processing unit 8 is unfolded into a rectangle in the circumferential direction to convert it into a planar unfolded captured image. For each concave and convex processing surface in the cylindrical hole portion 17a of the planar unfolded captured image, the brightness value is corrected and processed into a captured image with uniform brightness.

[0043] It should be noted that although the circumferential lengths of the various convex and concave surfaces are different, the difference is small, so images are taken using a plane that is approximately set to the same length.

[0044] In the third block P3, a partial image of the planar unfolded photograph is recorded for illustrative purposes.

[0045] In the fourth block P4, the image scale correction unit 9 divides the partial image of the planar unfolded image along the cylindrical direction according to each concave and convex processing surface. Based on the size information of each concave and convex processing surface, the resolution (pixels) of the aperture lens camera 2, and the grid pattern K (calibration image) of the captured image, the image scale correction value is calculated according to each concave and convex processing surface so that the captured image of each concave and convex processing surface becomes a uniform scale.

[0046] For example, the image scale correction values ​​for each concave and convex machined surface are: chamfered machined surface 17a1: 0.8, upper cylinder machined surface 17a2: 1.0, groove machined surface 17a3: 1.2, groove machined side surface 17a4: 0.8, and lower cylinder machined surface 17a5: 1.4.

[0047] In the fifth block P5, the image scale correction unit 9 generates a uniform scale segmentation scale correction image for each processing surface based on the image scale correction value of each processing surface.

[0048] In the sixth block P6, the image scale correction unit 9 combines the segmented scale correction images of each processing surface to generate a scale correction image.

[0049] This prevents defects displayed due to different scales from being missed, thus improving the accuracy of defect extraction.

[0050] Furthermore, since images with a uniform scale are obtained, it becomes possible to extract defects that take into account their size.

[0051] Furthermore, since it can obtain images with uniform brightness, it can prevent overexposure or underexposure that would make defects invisible and thus miss them, thereby improving the accuracy of defect extraction.

[0052] Figure 5 This is a timing diagram illustrating the generation process of the virtual boundary line image in Implementation Method 1. Figure 6 This is a timing diagram illustrating the process of obtaining the boundary line of a planar unfolded image of a rectangular image unfolded in the circumferential direction in Embodiment 1.

[0053] In the first block Q1, the cylindrical hole and camera coordinate system are configured in a virtual space such as 3DCAD and CG software, which are identical to the actual environment.

[0054] It should be noted that the relative positional relationship between the cylindrical hole and the camera coordinate system in the actual environment is measured by using calibration methods such as checkerboard patterns.

[0055] In the second block Q2, based on the camera's internal parameters (focal length, lens distortion, image pixel discrete values, etc.), a virtual image is generated that adds the virtual processing surface boundary to the virtual image of the cylindrical hole of the object observed from the camera coordinate system.

[0056] That is, in 3D space, for example by simply displaying the lines representing the boundaries of the virtual machining surfaces and projecting them onto the virtual captured image, it is possible to generate a virtual captured image with the extracted boundaries of the virtual machining surfaces appended to it.

[0057] Therefore, it is possible to geometrically extract the boundary line positions between the concave and convex surfaces within a virtual image.

[0058] In the third block Q3, the virtual image with the added virtual processing surface boundary is expanded into a rectangle in the circumferential direction.

[0059] In block Q4, for the image captured by the planar unfolding of a rectangle in the circumferential direction, the edge extraction method is used to estimate the candidate boundary lines of each processing surface.

[0060] Therefore, it is possible to deduce the boundary lines between the concave and convex surfaces in the unfolded image.

[0061] In the fifth block Q5, candidate line 1 is determined as the processing surface boundary line of the plane unfolded image corresponding to the boundary of the virtual processing surface of the detection object of the virtual captured image, based on the minimum distance between the boundary lines.

[0062] In the sixth block Q6, the processing of the fifth block Q5 is repeated to determine and obtain all the processing surface boundary lines of the corresponding planar unfolded image.

[0063] Therefore, it is possible to determine the boundary lines between the concave and convex surfaces in all planar unfolded images.

[0064] Figure 7 (a) and (b) are diagrams illustrating the method for determining that an image capture has been obstructed in Implementation 1.

[0065] That is, it is determined that there are un-captured (obscured) parts in the captured image.

[0066] (a) indicates a method for determining the occlusion (masking) of each processing surface that produced the captured image by using cross-determination calculation of virtual images of each processing surface and the optical axis of the camera, and the part that can be determined to be occluded (masked) is a part of the groove processing surface 17a3 and the groove processing side surface 17a4.

[0067] (b) indicates a method for determining the occluded (covered) portions of each processed surface in the captured image when the boundary line of the captured image corresponding to the boundary line of the virtual processed surface of the detection object in the virtual image is obtained, based on... Figure 6 In the fifth block Q5, where the minimum distance between the boundary lines determines the candidate lines as the boundary lines of the processing surface in the planar unfolded image corresponding to the boundary of the virtual processing surface of the detection object, if there are no processing surface boundary lines in the planar unfolded image corresponding to the boundary of the virtual processing surface of the detection object within a specified threshold distance, it is determined that the processing surface is occluded (hidden) and the processing surface boundary lines cannot be seen.

[0068] Therefore, even if a part of the image is not visible due to occlusion, a uniform scale image can be obtained, making it possible to extract defects that take into account their size.

[0069] Figure 8 This is a flowchart illustrating the process of determining occlusion (masking) and calculating image scaling values ​​in Implementation 1.

[0070] In step S11, based on the size information of the drawing information and the resolution of the camera, a setting value is set for how many pixels the actual size inside the cylindrical hole (actual size) will be displayed in the captured image (e.g., 1mm = 5 pixels).

[0071] In step S12, it is determined whether there is no obstruction (masking) on ​​each processing surface of the captured image so that the entire surface is displayed in the captured image.

[0072] If there is no obstruction (coverage) on any of the processing surfaces of the captured image, proceed to step S13; if there is obstruction (coverage) on any of the processing surfaces of the captured image, proceed to step S16.

[0073] In step S13, the actual size and set value of the cylindrical width of each processed surface are used to calculate how many pixels the width will be in the image.

[0074] In step S14, the pixel width in the captured image is measured.

[0075] In step S15, the image scale correction value for each processing surface is calculated.

[0076] (For example, if the actual width of the machined surface in the cylindrical direction is 2mm and is displayed as 5 pixels, and you want to enlarge 2mm to 10 pixels for display, then the image scale correction value is 2.0.)

[0077] In step S16, the width of the grid in the cylindrical direction of each processed surface is measured in pixels in the captured image based on the grid pattern (calibration image) projected by laser or the like.

[0078] In step S17, the image scale correction value is calculated based on the actual size of the grid on each processed surface of the captured image and the set value / pixel width.

[0079] (For example, if a 1mm wide grid is displayed as 2 pixels in the cylindrical direction, and you want to correct it to 1mm = 5 pixels, then the image scale correction value is 5 / 2 = 2.5.)

[0080] Therefore, even if there are areas in the image that cannot be seen due to occlusion, it is possible to obtain an image with a uniform scale. By using a grid pattern K (calibration image) projected by laser or other means, the scale correction value of the areas that cannot be seen due to occlusion can be calculated, and an image with a uniform scale can be obtained. Thus, it becomes possible to extract defects that take into account the defect size.

[0081] Next, the effects of Implementation Method 1 will be explained.

[0082] (1) The image inspection device 1 includes: an image acquisition unit 7, which acquires an image obtained by the aperture lens camera 2 of the cylindrical hole portion 17a; a camera information storage unit 12, which stores camera distance information between the aperture lens camera 2 and the cylindrical hole portion 17a; a drawing information storage unit 13, which stores drawing information having size information of the concave and convex machined surfaces (17a1-17a5) in the cylindrical hole portion 17a; a planar unfolding processing unit 8, which uses the camera distance information and the drawing information to convert the image of the concave and convex machined surfaces (17a1-17a5) in the cylindrical hole portion 17a into a planar unfolded image; an image scale correction unit 9, which uses the camera distance information and the drawing information to calculate the image scale correction value according to each concave and convex machined surface (17a1-17a5) of the cylindrical hole portion 17a and generates a scale-corrected image of the planar unfolded image; and a defect extraction processing unit 10, which extracts defects from the scale-corrected image.

[0083] Therefore, it can prevent defects displayed due to different scales from being missed, and improve the accuracy of defect extraction.

[0084] (2) The image scale correction unit 9 generates a scale-corrected image that displays the concave and convex machined surfaces (17a1-17a5) inside the cylindrical hole 17a at a uniform scale.

[0085] Therefore, it is possible to obtain images with a uniform scale, thus making it possible to extract defects that take into account their size.

[0086] (3) The image is taken in one shot of the concave and convex machined surfaces inside the cylindrical hole 17a.

[0087] Therefore, it can shorten the shooting time.

[0088] (4) The image scale correction unit 9 determines, based on the camera distance information and the image information, that occlusion (masking) occurs in the captured image of the concave and convex machined surfaces (17a1-17a5) in the cylindrical hole portion 17a, and generates a scale correction image that is displayed at a uniform scale according to each concave and convex machined surface (17a1-17a5) in the cylindrical hole portion 17a, taking into account the occlusion.

[0089] Therefore, even if there are areas of the image that are partially obscured due to occlusion, it is possible to obtain images with a uniform scale, thus making it possible to extract defects that take into account their size.

[0090] (5) The image scale correction unit 9 calculates the image scale correction value according to the camera distance information, the map information, and the grid pattern K (calibration image) projected by laser or the like, according to each concave and convex machined surface (17a1-17a5) in the cylindrical hole 17a, and generates a scale correction image of the planar unfolded photographed image.

[0091] Therefore, it is possible to calculate the scale correction value of the area that is not visible due to occlusion (shading) and obtain a photographic image with a uniform scale, thus making it possible to extract defects that take into account the defect size.

[0092] (6) The planar unfolding processing unit 8 extracts the position of the virtual processing surface boundary line group between the concave and convex processing surfaces (17a1-17a5) in the virtual captured image within the cylindrical hole 17a, which is geometrically determined, based on the camera distance information and the image information.

[0093] Therefore, it is possible to geometrically extract the boundary line position between the concave and convex surfaces (17a1-17a5) within the virtual captured image.

[0094] (7) The planar unfolding processing unit 8 uses edge extraction processing on the captured image of the concave and convex machined surfaces (17a1-17a5) in the cylindrical hole portion 17a to estimate the boundary line between the concave and convex machined surfaces (17a1-17a5) in the cylindrical hole portion 17a in the captured image.

[0095] Therefore, the boundary line between the uneven surfaces (17a1-17a5) in the captured image can be deduced.

[0096] (8) The planar unfolding processing unit 8 calculates the distance between the boundary line group between the concave and convex machined surfaces (17a1-17a5) in the virtual captured image of the cylindrical hole portion 17a, which is geometrically determined, and the position of the boundary line group between the concave and convex machined surfaces (17a1-17a5) in the captured image, which is estimated in the captured image, based on the camera distance information and the image information. It also calculates the distance between the extracted virtual machined surface boundary line between the concave and convex machined surfaces (17a1-17a5) and the multiple boundary lines between the concave and convex machined surfaces (17a1-17a5) in the captured image, and determines which boundary line between the extracted virtual machined surface boundary line between the concave and convex machined surfaces (17a1-17a5) and the multiple boundary lines between the concave and convex machined surfaces (17a1-17a5) in the captured image corresponds to.

[0097] Therefore, it is possible to determine the boundary lines between the uneven surfaces (17a1-17a5) in all captured images.

[0098] (9) The image scale correction unit 9 divides the planar unfolded image according to the concave and convex processing surfaces (17a1-17a5) in the cylindrical hole 17a, and generates a scale-corrected image of the planar unfolded image divided according to the image scale correction value of each concave and convex processing surface (17a1-17a5) in the cylindrical hole 17a. Then, it combines the scale-corrected image of the divided planar unfolded image.

[0099] Therefore, it can prevent defects smaller than the specified size from being missed due to defects on the boundary line of the uneven machined surfaces (17a1-17a5) inside the cylindrical hole 17a being displayed as small-sized defects on the image, thereby improving the accuracy of defect extraction.

[0100] (10) The planar unfolding processing unit 8 converts the photographed image of the concave and convex machined surfaces (17a1-17a5) inside the cylindrical hole 17a into a planar unfolded photographed image by rectangular unfolding.

[0101] Therefore, distortion of the captured image that is displayed in a curved manner in the circumferential direction of the cylindrical hole 17a can be corrected, thus obtaining a captured image with a uniform scale in the circumferential direction of the cylindrical hole 17a, which can improve the accuracy of defect extraction.

[0102] (11) The image scale correction unit 9 corrects the brightness value for each concave and convex machined surface (17a1-17a5) in the cylindrical hole portion 17a of the planar unfolded image, and processes it into an image with uniform brightness. Based on the camera distance information and the image information, it calculates the image scale correction value for each concave and convex machined surface (17a1-17a5) in the cylindrical hole portion 17a and generates a scale-corrected image of the planar unfolded image.

[0103] This allows for the acquisition of images with uniform brightness, preventing overexposure and underexposure that would obscure defects, thus preventing missed detections and improving the accuracy of defect extraction.

[0104] [Implementation Method 2]

[0105] Figure 9 This is a schematic diagram of the overall image inspection device according to Embodiment 2. Figure 10 This is the main flowchart illustrating the image inspection process in Implementation Method 2.

[0106] In Embodiment 1, the image scale correction unit 9 calculates the image scale correction value according to each processing surface and generates a scale correction image, and the defect extraction processing unit 10 extracts defects from the scale correction image. However, in Embodiment 2, the image scale correction unit 9 calculates the image scale correction value according to each processing surface (…). Figure 10 In step S2a), the defect extraction processing unit 10 extracts defects from the captured image based on the image scale correction value. Figure 10 Step S3a).

[0107] The other structures are the same as those in Embodiment 1, therefore the same reference numerals are used to label the same structures and descriptions are omitted.

[0108] Therefore, in embodiment 2, it has the same effect as in embodiment 1.

[0109] [Other Implementation Methods]

[0110] The embodiments for implementing the present invention have been described above, but the specific structure of the present invention is not limited to the structure of the embodiments. Design changes and the like that that do not depart from the spirit of the invention are also included in the present invention.

[0111] The object to be inspected is not limited to VTC covers; any object with a narrow cylindrical or polygonal opening as the inspection area can be inspected by image examination of the opening.

[0112] It should be noted that the present invention is not limited to the embodiments described above, and includes various modifications. For example, the embodiments described above are detailed embodiments for the purpose of easily understanding and illustrating the present invention, and are not limited to having all the structures described. In addition, a part of the structure of a certain embodiment can be replaced with the structure of another embodiment, and it is also possible to add the structure of another embodiment to the structure of a certain embodiment. Furthermore, regarding a part of the structure of each embodiment, other structures can be added, deleted, or replaced.

[0113] This application claims priority based on Japanese Patent Application No. 2024-097854, filed on June 18, 2024. The entire disclosure of the specification, claims, drawings, and abstract of Japanese Patent Application No. 2024-097854, filed on June 18, 2024, is incorporated herein by reference in its entirety.

[0114] Explanation of reference numerals in the attached figures

[0115] 1: Image inspection device; 2: Hole lens camera (camera); 3: Computer; 4: CPU; 5: Memory; 7: Image acquisition unit; 8: Planar unfolding processing unit; 9: Image scale correction unit; 10: Defect extraction processing unit; 12: Camera information storage unit; 13: Drawing information storage unit; 17: VTC cover (object to be inspected); 17a: Cylindrical hole section (inspection area).

Claims

1. An image inspection apparatus for inspecting defects in an inspection area of ​​an object, comprising: The image acquisition unit acquires an image of the inspection area obtained by a camera; A camera information storage unit stores camera distance information between the camera and the inspection area; The drawing information storage unit stores drawing information containing dimensional information of the concave and convex machined surfaces of the inspection area; The planar unfolding processing unit uses the camera distance information and the image information to convert the captured image of the uneven processing surface of the inspection area into a planar unfolding captured image; The image scale correction unit uses the camera distance information and the image information to calculate the image scale correction value according to each concave and convex processing surface of the inspection area, and uses the image scale correction value to generate a scale-corrected image of the planar unfolded photographed image. as well as The defect extraction processing unit extracts defects from the scale-corrected image.

2. The image inspection apparatus according to claim 1, The image scale correction unit: A scale-corrected image is generated based on the uneven surfaces of the inspection area, displayed at a uniform scale.

3. The image inspection device according to claim 2, The image scale correction unit: Based on the camera distance information and the image information, it is determined whether occlusion occurs in the captured image of the uneven surface of the inspection area. If occlusion is determined to occur, the occlusion is taken into account and a scale-corrected image displayed at a uniform scale is generated according to each uneven surface of the inspection area.

4. The image inspection apparatus according to claim 3, The image scale correction unit: Based on the camera distance information, the image information, and the calibration image, the image scale correction value is calculated according to each concave and convex processing surface of the inspection area, and the image scale correction value is used to generate the scale correction image of the planar unfolded image.

5. The image inspection apparatus according to claim 1, The planar unfolding processing unit: Based on the camera distance information and the image information, the positions of the virtual processing surface boundary line groups between the concave and convex processing surfaces in the virtual captured image of the inspection area are extracted using geometric calculation.

6. The image inspection apparatus according to claim 5, The planar unfolding processing unit: Edge extraction processing is used on the captured images of the uneven surfaces of the inspection area to estimate the potential boundary lines between the uneven surfaces of the inspection area within the captured images.

7. The image inspection apparatus according to claim 6, The planar unfolding processing unit: Based on the extracted positions of the virtual machining surface boundary line group and the estimated positions of the candidate boundary lines, the distances between the extracted virtual machining surface boundary lines and the estimated multiple boundary lines are calculated. Determine which of the multiple estimated boundary lines the extracted virtual machining surface boundary line corresponds to.

8. The image inspection apparatus according to claim 1, The image scale correction unit: The planar unfolded image is segmented according to the concave and convex processing surfaces of the inspection area. Then, a scale-corrected image of the segmented planar unfolded image is generated using the image scale correction value according to the concave and convex processing surfaces of the inspection area. Finally, the scale-corrected image of the segmented planar unfolded image is combined with the scale-corrected image of the segmented planar unfolded image.

9. The image inspection apparatus according to claim 1, The planar unfolding processing unit: The image of the uneven surface of the inspection area is converted into a planar image by expanding the image into a rectangle.

10. The image inspection apparatus according to claim 1, The image scale correction unit: For each concave and convex surface of the planar unfolded image of the inspection area, the brightness value is corrected to process it into a uniformly bright image. Based on the camera distance information and the image information, the image scale correction value is calculated according to each concave and convex surface of the inspection area. The image scale correction value is used to generate a scale-corrected image of the planar unfolded image.

11. An image inspection apparatus for inspecting defects in an inspection area of ​​an object, comprising: The image acquisition unit acquires an image of the inspection area obtained by a camera; A camera information storage unit stores camera distance information between the camera and the inspection area; The drawing information storage unit stores drawing information containing dimensional information of the concave and convex machined surfaces of the inspection area; The planar unfolding processing unit uses the camera distance information and the image information to convert the captured image of the uneven processing surface of the inspection area into a planar unfolding captured image; The image scale correction unit uses the camera distance information and the image information to calculate the image scale correction value according to each concave and convex processing surface of the inspection area; as well as The defect extraction processing unit extracts defects from the images of the uneven surfaces of the inspection area, based on the image scale correction value and according to each uneven surface of the inspection area.

Images