Visual inspection device, visual inspection method, and control program for visual inspection device

Abstract

To provide a visual inspection device, a visual inspection method, and a control program of the visual inspection device, which can detect a defect of an inspection object with a large area in a shorter time than before.SOLUTION: A visual inspection device includes: striped pattern formation means for forming a striped pattern which comprises dark parts and bright parts parallel to each other on a surface of an inspection object; imaging means for taking an image of the surface of the inspection object; edge detection means for detecting an edge of the striped pattern in the image; characteristic amount calculation means for calculating the amount of characteristic of the edge; and defect presence determination means for determining the presence of a defect in the inspection object on the basis of the amount of the characteristic.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] This invention relates to a visual inspection device, etc. [Background technology]

[0002] Visual inspection devices for detecting defects in objects are used in various fields. In typical visual inspection devices, an image of the object's surface is first captured. Then, the image is analyzed to detect defects such as cracks. Various visual inspection methods have been proposed to detect true defects and prevent false positives.

[0003] For example, Patent Document 1 discloses a technology for an image processing device that extracts straight lines from an image in order to remove edges indicating non-linear cracks. In this image processing device, first, edges are detected from the image using a differential filter. Edges are detected in both the horizontal and vertical directions. Next, the pixels extracted as horizontal or vertical edges are subjected to a Hough transform. Then, based on the voting value which is the calculation result of the Hough transform, straight lines are extracted from the image. When this image processing device is applied to crack inspection of concrete walls, non-linear lines correspond to patterns that are cracks, etc., and straight lines correspond to patterns that are not cracks, etc. Therefore, by excluding straight lines in the image from the inspection target, the location of defects such as cracks can be identified.

[0004] Related technologies are also disclosed in Patent Documents 2 and 3. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2018-060343 [Patent Document 2] Japanese Patent Publication No. 2016-112947 [Patent Document 3] International Publication No. 2013 / 069100 [Overview of the Initiative] [Problems that the invention aims to solve]

[0006] The technology described in Patent Document 1 detects defects using an image of the defect itself. Therefore, detecting minute defects requires using an enlarged image of the target area. However, using enlarged images presents the problem of being time-consuming when inspecting large areas.

[0007] This invention has been made in view of the above-mentioned problems, and aims to provide an appearance inspection device, etc., that can detect defects in a large area of ​​an object to be inspected in a shorter amount of time. [Means for solving the problem]

[0008] To solve the above problems, the present invention provides a visual inspection apparatus comprising: stripe pattern forming means for forming a stripe pattern consisting of parallel dark and light areas on the surface of an object to be inspected; imaging means for capturing an image of the surface of the object to be inspected; edge detection means for detecting the edges of the stripe pattern in the image; feature quantity calculation means for calculating the feature quantities of the edges; and defect presence / absence determination means for determining the presence or absence of defects in the object to be inspected based on the feature quantities.

[0009] Furthermore, the present invention provides a visual inspection apparatus method in which the visual inspection apparatus forms a striped pattern consisting of parallel dark and light areas on the surface of the object to be inspected, captures an image of the surface of the object to be inspected, detects the edges of the striped pattern in the image, calculates the feature quantities of the edges, and determines whether or not there are defects in the object to be inspected based on the feature quantities.

[0010] In addition, the control program of the appearance inspection device of the present invention is a control program for an appearance inspection device. The appearance inspection device includes a striped pattern forming means for forming a striped pattern consisting of dark and bright parts parallel to each other on the surface of the inspection object, and an imaging means for imaging an image of the surface of the inspection object. The appearance inspection device is caused to execute a process of detecting an edge of the striped pattern in the image, a process of calculating a feature amount of the edge, and a process of determining the presence or absence of a defect in the inspection object based on the feature amount.

Effect of the Invention

[0011] The effect of the present invention is to provide an appearance inspection device or the like that can detect defects in an inspection object with a wide area in a shorter time.

Brief Description of the Drawings

[0012] [Figure 1] It is a block diagram showing an appearance inspection device of the first embodiment. [Figure 2] It is a schematic diagram showing an example of a striped pattern in an image of the first embodiment. [Figure 3] It is a schematic diagram showing another example of a striped pattern in an image of the first embodiment. [Figure 4] It is a block diagram showing a feature amount calculation means of the first embodiment. [Figure 5] It is a schematic diagram showing an example of a feature amount of the first embodiment. [Figure 6] It is a schematic diagram for explaining a distance calculation method of a feature amount of the first embodiment. [Figure 7] It is a block diagram showing a defect presence / absence determination means of the first embodiment. [Figure 8] It is a graph showing an example of a feature amount distribution of the first embodiment. [Figure 9] It is a graph showing another example of a feature amount distribution of the first embodiment. [Figure 10] It is a flowchart showing the operation of the appearance inspection device of the first embodiment. [Figure 11]This is a schematic diagram showing a specific example of the configuration of the visual inspection device according to the first embodiment. [Modes for carrying out the invention]

[0013] Embodiments of the present invention will be described in detail below with reference to the drawings. However, the embodiments described below include technically preferred limitations for carrying out the present invention, but do not limit the scope of the invention. Similar components in each drawing are given the same number and their descriptions may be omitted.

[0014] (First Embodiment) Figure 1 is a block diagram showing a visual inspection apparatus 100 according to the first embodiment. The visual inspection apparatus includes a striped pattern forming means 10, an imaging means 20, an edge detection means 30, a feature quantity calculation means 40, and a defect presence / absence determination means 50.

[0015] The striped pattern forming means 10 forms a striped pattern on the surface of the object to be inspected. The striped pattern consists of parallel dark and light areas.

[0016] The imaging means 20 captures an image of the surface of the object to be inspected.

[0017] The edge detection means 30 detects the edges of the striped pattern in the captured image. For example, the edge detection means 30 detects edges using the Canny method. In the Canny method, first, a brightness gradient is obtained in an image in which multiple pixels are arranged. Then, the pixel where the brightness gradient is maximum is extracted as an edge. At this time, the direction of the brightness gradient is detected along with the edge. In this embodiment, the vector indicating the direction of the brightness gradient is called the edge normal vector.

[0018] The feature calculation means 40 calculates the edge features.

[0019] The defect detection means 50 determines whether or not there are defects in the object to be inspected. The determination is made based on the acquired feature quantities.

[0020] Next, we will explain the details using a specific example. Figure 2 is a schematic diagram showing an example of a striped pattern in an image of the first embodiment. The surface of the object to be inspected 90 is a uniform plane. When the surface is a uniform plane, the striped pattern forming means 10 forms a striped pattern 200 on the object to be inspected 90 in which striped dark areas 201 and light areas 202 are arranged alternately and parallel to each other.

[0021] Figure 3 is a schematic diagram showing another example of the striped pattern in the image of the first embodiment. On the other hand, if there is a defect 91 in the object to be inspected 90, distortion occurs in the striped pattern 200 formed on the surface of the object to be inspected 90 by the striped pattern forming means 10. The visual inspection device 100 uses this distortion to determine whether or not there is a defect 91.

[0022] Figure 4 is a block diagram showing the feature calculation means 40 of the first embodiment. The feature calculation means 40 includes a normal vector calculation means 41, a reference line calculation means 42, and a distance calculation means 43.

[0023] The normal vector calculation means 41 calculates the normal vector of the edge. In the Canny method described above, the edge detection means 30 detects the direction of the brightness gradient. The normal vector of the edge indicates the direction of the brightness gradient. Based on this direction of the brightness gradient, the normal vector calculation means 41 calculates the normal vector of the edge. The normal vector of the edge is calculated for each pixel detected as an edge. In the field of image processing, the normal of a contour is the normal described above. The normal vector represents the direction of this normal.

[0024] The reference line calculation means 42 calculates a reference line. The reference line is a straight line passing through the origin of the image. For example, the top left of the image is considered the origin. The reference line is parallel to the normal vector of the edge. A reference line is calculated for each pixel detected as an edge.

[0025] The distance calculation means 43 calculates the distance from the reference line to the edge. The distance is calculated for each pixel detected as an edge.

[0026] The feature quantity calculation means 40 calculates the direction and distance of the normal vector of the above-mentioned edge as the feature quantity of the edge. The normal vector is a unit vector with a length of 1.

[0027] Next, the direction and distance of the normal vector will be described using a specific example. FIG. 5 is a schematic diagram showing an example of the feature quantity of the first embodiment. The feature quantity is the direction θ of the normal vector of the edge and the distance d. Here, the coordinates in the image are represented by (u, v). In FIG. 5, the upper left is the origin O, the right direction is the u-axis, and the lower direction is the v-axis. The direction θ is the angle clockwise from the u-axis. Also, the coordinates of each pixel are measured at the center of the pixel.

[0028] In FIG. 5, a state where both sides of the dark part 201 are detected as edges is shown. That is, the direction in which the luminance increases from bottom to top is positive.

[0029] In FIG. 5, the directions of the normal vectors of the first to third rows counted from the top are the same (θ1). The directions of the normal vectors of the fourth to sixth rows are θ2. Also, θ1 and θ2 are different. In FIG. 5, the former and latter normal vectors n are represented as n (u1,v1) 、n (u2,v2) . Here, n (u,v) indicates that n is a function of u and v.

[0030] Next, the calculation method of the distance d will be described. FIG. 6 is a schematic diagram showing the distance calculation method of the feature quantity of the first embodiment. The coordinates of the pixel are (u, v). At this time, there is an angular relationship as shown in FIG. 6 between the reference line L (u,v) , the line segment connecting the origin and the edge pixel, and the perpendicular line dropped from the edge pixel to the reference line L (u,v) . Using this relationship, the distance d (u,v) is calculated by the following formula.

[0031] d (u,v) =(u 2 +v 2 ) 1 / 2 ·sin{(π - θ)+tan -1(v / u)} The d calculated above (u,v) and the direction of the normal vector θ (u,v) These represent the feature quantities of the target edge pixels.

[0032] By following the procedure described above, for each edge pixel, the feature quantity θ (u,v) and d (u,v) The feature calculation means 40 calculates each of these.

[0033] Next, a method for determining the presence or absence of defects will be described. Figure 7 is a block diagram of the defect presence / absence determination means 50 of the first embodiment. The defect presence / absence determination means 50 comprises a feature distribution generation means 51, a cluster generation means 52, and a determination means 53.

[0034] The feature distribution generation means 51 generates a feature distribution. In the following explanation, θ (u,v) and d (u,v) These are simply abbreviated as θ and d. Prior to generating the feature distribution, the feature distribution generation means 51 performs a process to align the direction of the normal vector by replacing θ with θ+π when θ is negative. Next, the feature quantities of each edge pixel are plotted in a feature space with θ and d as axes.

[0035] The cluster generation means 52 groups multiple edges plotted in the feature space. This grouping allows the cluster generation means 52 to generate clusters. In cluster generation, multiple edges whose positional relationship in the feature space falls within a first range are grouped. Further details will be described later.

[0036] The determination means 53 determines whether or not there are defects in the inspection target 90 based on the number of clusters. Based on the number of faces of the inspection target 90, if the number of clusters is greater than the number of faces, the determination means 53 determines that there are defects. On the other hand, if the number of clusters is the same as the number of faces, the determination means 53 determines that there are no defects. Here, the number of faces is the number of faces with different inclinations from the reference face.

[0037] Next, we will explain how to determine the presence or absence of defects using a specific example. Figure 8 is a graph showing an example of the feature distribution of the first embodiment. The feature quantities of each edge are plotted in the feature space. In this example, there are no defects 91 in the object to be inspected 90. Therefore, the surface of the object to be inspected 90 is a uniform plane. Under the above conditions, most of the normal vectors are in the same direction. Therefore, the plotted edges are concentrated in the vicinity of the same θ1. Note that in visual inspection, there are measurement errors and quantization errors. Therefore, the θ of each edge does not necessarily coincide with θ1. In other words, the θ of each edge is distributed with a certain range. Therefore, taking this distribution into consideration, a first range for θ in the direction is determined. Also, the distance d, by its definition, can take a range approximately equal to the width of the image. Therefore, a first range for d in the distance is determined. The first range for the positional relationship is determined by combining the first range for the direction and the first range for the distance.

[0038] The cluster generation means 52 groups edges whose positional relationship falls within the first range described above. This grouping allows the cluster generation means 52 to generate clusters. In Figure 8, one cluster C1 is generated. Figure 8 also shows scattered edges other than cluster C1, but these are located further away from the first range in terms of θ or d. Therefore, these points do not form clusters. For example, these points are considered noise. If the surface of the object to be inspected 90 is free of cracks and uniform, then only one cluster is formed. Here, a crack is defined as the presence of multiple surfaces with different inclinations from the reference plane. It is believed that defects such as cracks 91 cause these cracks. The presence of multiple clusters means that there are surfaces with different inclinations corresponding to the number of clusters.

[0039] Based on the causal relationship described above, if the number of clusters is greater than the number of faces with different inclinations from the reference plane, then there is a defect 91. In the following explanation, the number of faces with different inclinations from the reference plane will be referred to as the number of faces. For the reasons stated above, the determination means 53 compares the number of clusters with the number of faces to determine whether or not there is a defect. If the surface of the object to be inspected 90 is a single plane, the number of faces is 1. Therefore, if only one cluster is generated, the determination means 53 determines that there is no defect. On the other hand, if two or more clusters are formed, the determination means 53 determines that there is a defect.

[0040] Similarly, for an inspection target 90 with two faces, if the number of clusters is two, the determination means 53 determines that there are no defects. On the other hand, if the number of clusters is three or more, the determination means 53 determines that there are defects. Similarly, if the inspection target 90 has three or more faces, the determination means 53 determines whether or not there are defects. Note that even if the number of faces is two or more, if the inspection range is narrowed to just one face, it becomes the same as an inspection target 90 with one face.

[0041] Figure 9 is a graph showing another example of the feature distribution in the first embodiment. Similar to the example in Figure 8, the number of faces of the object to be inspected 90 is 1. However, in the example in Figure 9, two clusters are generated. Specifically, cluster C2 is generated near θ2, and cluster C3 is generated near θ3. In other words, there are 2 clusters for 1 face. For this reason, the determination means 53 determines that there is a defect.

[0042] Based on the above explanation, the operation of the visual inspection device 100 will now be described. Figure 10 is a flowchart showing the operation of the visual inspection device 100 according to the first embodiment. First, the striped pattern forming means 10 forms a striped pattern 200 on the surface of the object to be inspected 90 (S1). Next, the imaging means 20 captures an image of the surface of the object to be inspected 90 (S2). Next, the edge detection means 30 detects the edges of the striped pattern 200 (S3). At this time, edges are detected on a pixel-by-pixel basis. The direction of the brightness gradient is also detected simultaneously. Next, the feature calculation means 40 calculates the feature quantity of each edge (S4). Next, the feature quantity distribution generation means 51 generates a feature quantity distribution (S5). This distribution is generated by plotting the feature quantity of each edge in the feature space. Next, the cluster generation means 52 generates clusters (S6). Clusters are generated by grouping multiple edges whose positional relationship is within the first range. Finally, the determination means 53 determines whether or not there is a defect (S7). The determination means 53 compares the number of faces of the 90 objects to be inspected with the number of clusters generated to determine whether or not there are defects.

[0043] Next, a specific example of the configuration of the visual inspection device 100 will be described. Figure 11 is a schematic diagram showing a specific example of the configuration of the visual inspection device 100 according to the first embodiment. In this configuration example, the visual inspection device 100 forms a striped pattern 200 through interference. This utilizes the configuration of a so-called Fizeau interferometer.

[0044] The visual inspection device 100 includes a striped pattern forming means 10 consisting of a light source 11, a beam splitter 12, a collimator lens 13, a reference plate 14, and a stage 15. The visual inspection device 100 also includes an imaging means 20 consisting of a lens 21 and an image sensor 22. An edge detection means 30, a feature quantity calculation means 40, and a defect presence / absence determination means 50 are implemented in the computer 60. The computer 60 includes a processor, memory, storage unit, input / output devices, etc.

[0045] Next, the method for forming the striped pattern 200 will be described. Light emitted from the light source 11 has its optical path changed by the beam splitter 12. Then, the incident light is converted into parallel light by the collimator lens 13. The reference plate 14 reflects some of the light and transmits some of the light. A reference surface 14a that reflects light is formed on the back surface of the reference plate 14. Light that has passed through the reference plate 14 is reflected by the surface of the object to be inspected 90. The stage 15 that holds the object to be inspected 90 is equipped with a tilt adjustment mechanism. If the reference surface 14a and the surface of the object to be inspected 90 are not parallel, the reflected light from the reference surface 14a and the reflected light from the surface of the object to be inspected 90 interfere with each other, generating interference fringes. These interference fringes become the striped pattern 200. The spacing of the interference fringes can be adjusted by adjusting the tilt of the stage 15. Note that in Figure 11, the tilt of the stage 15 is visible for illustrative purposes. However, the tilt adjustment is a gap adjustment on the order of the wavelength of visible light. Therefore, in reality, the reference surface 14a and the surface of the object being inspected 90 appear almost parallel to the naked eye.

[0046] Interfering light emitted from the reference plate 14 passes through the collimator lens 13, beam splitter 12, and lens 21. The interfering light then enters the image sensor 22. As a result, an image of the inspection target 90 with the striped pattern 200 superimposed is captured.

[0047] The captured image is analyzed by the edge detection means 30, feature calculation means 40, and defect presence / absence determination means 50 implemented in the computer 60. These analysis means determine whether or not there are defects in the inspection target 90.

[0048] In the above example, the striped pattern 200 was formed by a Fizeau interferometer. However, the striped pattern 200 may also be formed by light interference caused by, for example, monochromatic light being incident on a diffraction grating. Alternatively, the striped pattern 200 may be formed by, for example, projecting the pattern that will become the striped pattern 200 onto the object to be inspected 90.

[0049] As described above, with the visual inspection device 100 of this embodiment, a striped pattern 200 is formed on the surface of the object to be inspected 90, and the image is analyzed. Pixels constituting the edges of the striped pattern 200 are detected. Then, by analyzing the feature quantities of these edges, the presence or absence of defects in the object to be inspected 90 is determined. In this determination, it is not necessary to take an enlarged image to detect the defect itself. Therefore, it becomes possible to detect defects in a large area of ​​the object to be inspected 90 in a shorter time. By analyzing the image of the object to be inspected 90 with the superimposed pattern, the location of defects on the surface of the object to be inspected 90 can be estimated. In this case, it is not necessary to acquire an enlarged image of the object to be inspected 90. Therefore, it becomes possible to detect defects in a large area of ​​the object to be inspected 90 more quickly.

[0050] The visual inspection device 100 and other components of this embodiment have been described above.

[0051] The visual inspection apparatus 100 of this embodiment includes a striped pattern forming means 10, an imaging means 20, an edge detection means 30, a feature quantity calculation means 40, and a defect presence / absence determination means 50. The striped pattern forming means 10 forms a striped pattern 200 consisting of parallel dark areas 201 and bright areas 202 on the surface of the object to be inspected 90. The imaging means 20 captures an image of the surface of the object to be inspected 90. The edge detection means 30 detects the edges of the striped pattern 200 in the image. The feature quantity calculation means 40 calculates the feature quantities of the edges. The defect presence / absence determination means 50 determines whether or not there are defects in the object to be inspected 90 based on the feature quantities.

[0052] In the above configuration, the presence or absence of defects can be determined using an overall image of the inspection target 90, including the striped pattern 200. In the above configuration, the edge features of the captured striped pattern 200 are utilized. Furthermore, the presence or absence of defects is determined based on the distribution of these features. Therefore, there is no need to capture magnified images to directly detect defects 91. As a result, it becomes possible to detect defects in a large area of ​​the inspection target 90 in a shorter amount of time.

[0053] In another embodiment, the feature quantity calculation means 40 of the visual inspection device 100 comprises a normal vector calculation means 41, a reference line calculation means 42, and a distance calculation means 43. The normal vector calculation means 41 calculates a normal vector indicating the direction of the edge normal. The reference line calculation means 42 calculates a reference line that passes through the origin of the image and is parallel to the normal vector. The distance calculation means calculates the distance from the reference line to the edge. The feature quantity calculation means 40 then acquires the direction of the normal vector and the distance as feature quantities.

[0054] In this configuration, the direction θ of the normal vector and the distance d are obtained as edge features.

[0055] In another embodiment, the defect determination means of the visual inspection device 100 comprises a feature distribution generation means 51, a cluster generation means 52, and a determination means 53. The feature distribution generation means 51 generates a feature distribution by plotting the feature quantities of edges in the feature space. The cluster generation means 52 generates clusters by grouping multiple edges whose positional relationship in the feature space is within a first range. When the number of clusters is greater than the number of faces of the inspection target 90, the determination means 53 determines that there is a defect in the inspection target 90.

[0056] In the above configuration, the presence or absence of defects is determined by analyzing an overall image of the inspection target 90. In other words, there is no need to take magnified images to directly detect defects 91. Therefore, it becomes possible to detect defects on the surface of the inspection target 90, which has a large area, in a shorter amount of time.

[0057] In another embodiment, the striped pattern forming means 10 of the visual inspection device 100 forms a striped pattern 200 by light interference. In this process, interference fringes are formed by the interference of reflected light from the surface of the object to be inspected 90 and reflected light from the reference surface. With the above configuration, a striped pattern 200 consisting of parallel dark areas 201 and light areas 202 can be easily formed.

[0058] In another embodiment, in the visual inspection device 100, the striped pattern forming means 10 forms a striped pattern 200 by interference of monochromatic light.

[0059] With the above configuration, a striped pattern 200 consisting of parallel dark areas 201 and light areas 202 can be easily formed.

[0060] In another embodiment, in the visual inspection device 100, the striped pattern forming means 10 forms a striped pattern 200 by projection of a pattern. With the above configuration, a striped pattern 200 consisting of parallel dark areas 201 and light areas 202 can be easily formed.

[0061] In the visual inspection method of this embodiment, the visual inspection device 100 forms a striped pattern 200 on the surface of the object to be inspected 90. The striped pattern 200 consists of parallel dark areas 201 and light areas 202. The visual inspection device 100 also captures an image of the surface of the object to be inspected 90. The visual inspection device 100 also detects the edges of the striped pattern 200 in the image. The visual inspection device 100 also calculates feature quantities for the edges. Based on these feature quantities, the visual inspection device 100 determines whether or not there are defects in the object to be inspected 90.

[0062] In the above configuration, the presence or absence of defects can be determined using an overall image of the inspection target 90, including the striped pattern 200. In the above configuration, the edge features of the captured striped pattern 200 are utilized. Furthermore, the presence or absence of defects is determined based on the distribution of these features. Therefore, there is no need to take magnified images to directly detect defects 91. As a result, it becomes possible to detect defects on the surface of a large area of ​​the inspection target 90 in a shorter amount of time.

[0063] In another embodiment, in the visual inspection method, the visual inspection device 100 calculates a normal vector. The normal vector indicates the direction of the edge's normal. The visual inspection device 100 also calculates a reference line. The reference line passes through the origin of the image and is parallel to the normal vector. The visual inspection device 100 also calculates the distance from the reference line to the edge. The visual inspection device 100 then acquires the direction of the normal vector and the distance as feature quantities.

[0064] In this configuration, the direction θ of the normal vector and the distance d are obtained as edge features.

[0065] Furthermore, the control program for the visual inspection apparatus of this embodiment controls the visual inspection apparatus 100. Here, the visual inspection apparatus 100 comprises a striped pattern forming means 10 and an imaging means 20. The striped pattern forming means 10 forms a striped pattern 200 on the surface of the object to be inspected 90. The striped pattern 200 consists of parallel dark areas 201 and bright areas 202. The control program for the visual inspection apparatus causes the visual inspection apparatus 100 to perform the following processes. This process is to detect the edges of the striped pattern in the image. This process is to calculate the feature quantities of the edges. This process is to determine whether or not there are defects in the object to be inspected 90 based on the feature quantities.

[0066] According to the control program of this visual inspection device, the presence or absence of defects can be determined using an overall image of the inspection target 90, including the striped pattern 200. In the above configuration, the edge features of the captured striped pattern 200 are used. Furthermore, the presence or absence of defects is determined based on the distribution of these features. Therefore, there is no need to take magnified images to directly detect defects 91. As a result, it becomes possible to detect defects in a large area of ​​the inspection target 90 in a shorter amount of time.

[0067] The present invention also includes a program that causes a computer to execute the processing of the embodiments described above, and a recording medium storing said program. Examples of recording media that can be used include magnetic disks, magnetic tapes, optical disks, magneto-optical disks, semiconductor memory, and the like.

[0068] The present invention has been described above using the embodiments as exemplary examples. However, the present invention is not limited to the embodiments described above. That is, the present invention can be applied in various forms that can be understood by those skilled in the art, within the scope of the present invention. [Explanation of symbols]

[0069] 10. Striped pattern forming means 20 Imaging means 30 Edge detection means 40 Feature Calculation Method 41 Normal vector calculation means 42. Reference line calculation means 43 Distance Calculation Means 50 Defect detection method 51 Feature distribution generation means 52 Cluster generation means 53 Judgment means 60 Computer 90 subjects of testing 100 Visual Inspection Device

Claims

1. A striped pattern forming means for forming a striped pattern consisting of parallel dark and light areas on the surface of the object to be inspected, An imaging means for capturing an image of the surface of the object to be inspected, An edge detection means for detecting the edges of the striped pattern in the aforementioned image, A feature calculation means for calculating the feature quantities of the aforementioned edge, A defect presence / absence determination means for determining the presence or absence of defects in the object to be inspected based on the aforementioned feature quantities, It has, The feature calculation means, A normal vector calculation means for calculating a normal vector indicating the direction of the normal of the edge, A reference line calculation means for calculating a reference line that passes through the origin of the aforementioned image and is parallel to the aforementioned normal vector, Distance calculation means for calculating the distance from the reference line to the edge, Equipped with, The direction of the normal vector and the distance are obtained as the feature quantities. A visual inspection device characterized by the following features.

2. A striped pattern forming means for forming a striped pattern consisting of parallel dark and light areas on the surface of an object to be inspected, An imaging means for capturing an image of the surface of the object to be inspected, An edge detection means for detecting the edges of the striped pattern in the aforementioned image, A feature calculation means for calculating the feature quantities of the aforementioned edge, A defect presence / absence determination means for determining the presence or absence of defects in the object to be inspected based on the aforementioned feature quantities, It has, The defect determination means is A feature distribution generation means that plots the feature quantities of the edge in the feature space to generate a feature distribution, A cluster generation means that groups a plurality of edges whose positional relationship in the feature space is within a first range to generate a cluster, A determination means for determining that there is a defect in the object to be inspected when the number of clusters is greater than the number of surfaces to be inspected, An appearance inspection device characterized by having the following features.

3. The feature calculation means, A normal vector calculation means for calculating a normal vector indicating the direction of the normal of the edge, A reference line calculation means for calculating a reference line that passes through the origin of the aforementioned image and is parallel to the aforementioned normal vector, Distance calculation means for calculating the distance from the reference line to the edge, Equipped with, The direction of the normal vector and the distance are obtained as the feature quantities. The visual inspection apparatus according to feature 2.

4. The striped pattern forming means is The aforementioned striped pattern is formed by the interference of reflected light from the surface of the object to be inspected and reflected light from the reference surface. The appearance inspection apparatus according to any one of claims 1 to 3.

5. The striped pattern forming means is The aforementioned striped pattern is formed by the interference of monochromatic light. The visual inspection apparatus according to feature 1.

6. The striped pattern forming means is The aforementioned striped pattern is formed by projection of the pattern. The visual inspection apparatus according to feature 1.

7. The visual inspection device, A striped pattern consisting of parallel dark and light areas is formed on the surface of the object to be inspected. An image of the surface of the object to be inspected is captured, The edges of the striped pattern in the aforementioned image are detected, The normal vector indicating the direction of the normal of the aforementioned edge is calculated, A reference line passing through the origin of the aforementioned image and parallel to the aforementioned normal vector is calculated, The distance from the reference line to the edge is calculated, The direction of the normal vector and the distance are obtained as feature quantities. Based on the aforementioned feature quantities, the presence or absence of defects in the object to be inspected is determined. A visual inspection method characterized by the following features.

8. The visual inspection device is A striped pattern consisting of parallel dark and light areas is formed on the surface of the object to be inspected. An image of the surface of the object to be inspected is captured, The edges of the striped pattern in the aforementioned image are detected, The feature quantities of the aforementioned edge are calculated, Based on the aforementioned feature quantities, the presence or absence of defects in the object to be inspected is determined. A visual inspection method characterized by the following: The feature quantities of the edge are plotted in the feature space to generate a feature distribution. Multiple edges whose positional relationship in the feature space is within a first range are grouped to generate clusters. When the number of clusters is greater than the number of surfaces to be inspected, it is determined that there is a defect in the surface to be inspected. A visual inspection method characterized by the following features.

9. A control program for a visual inspection device, The aforementioned visual inspection device is, A striped pattern forming means for forming a striped pattern consisting of parallel dark and light areas on the surface of the object to be inspected, An imaging means for capturing an image of the surface of the object to be inspected, Equipped with, The aforementioned visual inspection device, A process for detecting the edges of the striped pattern in the aforementioned image, A process for calculating a normal vector that indicates the direction of the normal of the edge, A process to calculate a reference line that passes through the origin of the aforementioned image and is parallel to the aforementioned normal vector, A process for calculating the distance from the aforementioned reference line to the aforementioned edge, A process to obtain the direction of the normal vector and the distance as feature quantities of the edge, A process to determine whether or not there are defects in the object to be inspected based on the aforementioned feature quantities, A control program for a visual inspection device, characterized by causing it to execute.

10. A control program for a visual inspection device, The aforementioned visual inspection device is, A striped pattern forming means for forming a striped pattern consisting of parallel dark and light areas on the surface of the object to be inspected, An imaging means for capturing an image of the surface of the object to be inspected, Equipped with, The aforementioned visual inspection device, A process for detecting the edges of the striped pattern in the aforementioned image, The process of calculating the feature quantities of the aforementioned edge, The process involves plotting the aforementioned features in a feature space to generate a feature distribution, A process of grouping multiple edges whose positional relationship in the feature space is within a first range to generate clusters, A process to determine that there is a defect in the object to be inspected when the number of clusters is greater than the number of surfaces to be inspected, A control program for a visual inspection device, characterized by causing it to execute.

Images