Visual inspection device and visual inspection method

Abstract

To provide a visual inspection device that can acquire information for appropriately detecting a defective part of an appearance image in which an object is captured even when the degree of surface roughness varies depending on the object.SOLUTION: A visual inspection device 100 comprises: an image acquisition unit 5 that acquires an appearance image 10 of a semiconductor wafer U; and a processing unit 1 that performs filter processing for detection on the appearance image 10 to create an image for detection 10c for detecting a defective part 12. The processing unit 1 performs the filter processing for detection so as to decrease detection sensitivity for the defective part 12 and a defective part 13 as the degree of roughness of the surface S of the semiconductor wafer U captured in the appearance image 10 increases, and thereby creates the image for detection 10c according to the degree of roughness of the surface S of the semiconductor wafer U captured in the appearance image 10.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an appearance inspection device and an appearance inspection method for detecting defects in an object shown in an appearance image. [Background technology]

[0002] 2. Description of the Related Art Conventionally, there has been known an appearance inspection device (defect inspection device) for detecting defects in an object shown in an appearance image (see, for example, Patent Document 1).

[0003] The above-mentioned Patent Document 1 discloses a defect inspection device that inspects the presence or absence of defects in a wafer based on an image of the wafer, which is an object of the inspection. In the defect inspection device of the above-mentioned Patent Document 1, when an image of the wafer captured by an imaging device (image acquisition unit) is input to a control device (processing unit), a defective area existing on the surface of the wafer is identified by a program of the control device. Specifically, in the defect inspection device of the above-mentioned Patent Document 1, the control device digitizes the captured image as pixel values ​​on a pixel-by-pixel basis, and identifies pixels whose pixel values ​​exceed a certain threshold value as defective areas. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2014-115245 A Summary of the Invention [Problem to be solved by the invention]

[0005] In the defect inspection device of the above-mentioned Patent Document 1, the control device digitizes the captured image as pixel values ​​in pixel units, and identifies pixels whose pixel values ​​exceed a certain threshold value as defective areas where defects exist. However, the wafer, which is the object of the defect inspection, may have a rough surface portion, which is a portion where the pixel value is not constant, even though there is no defect on the surface. Furthermore, the degree of surface roughness may differ from wafer to wafer, and when the presence or absence of a defective area is judged based on a certain threshold value for each of the appearance images in which wafers with different degrees of surface roughness are captured, information for appropriately detecting the defective area cannot be obtained, and erroneous detection or detection failure may occur. Specifically, when the detection sensitivity of the defective area is increased in the defect inspection, the surface roughness may be erroneously detected as a defective area in a wafer with a large degree of surface roughness, even in a portion where there is no defect such as dirt. In addition, when the detection sensitivity of the defective area is decreased in the defect inspection, the defective area may not be detected in a wafer with a small degree of surface roughness, even if there is a defect. Therefore, there is a demand for an appearance inspection device that can obtain information for appropriately detecting a defective part in an appearance image in which an object is captured, even when the degree of surface roughness differs depending on the object.

[0006] The present invention has been made to solve the above-mentioned problems, and one object of the present invention is to provide an appearance inspection device that can acquire information for properly detecting defective areas in an appearance image showing an object, even when the degree of surface roughness varies depending on the object. [Means for solving the problem]

[0007] In order to achieve the above object, the appearance inspection device according to the first aspect includes an image acquisition unit that acquires an appearance image of an object, and a processing unit that performs detection filter processing on the appearance image to generate a detection image for detecting a defect part contained in the appearance image, and the processing unit is configured to perform detection filter processing such that the detection sensitivity to the defect part is reduced as the degree of roughness of the surface of the object captured in the appearance image increases, thereby generating a detection image according to the degree of roughness of the surface of the object captured in the appearance image. Here, the "defect part" of the present invention is not limited to partial defects and scratches of the object captured in the appearance image, but is a broad concept including dirt and foreign matter adhesion.

[0008] In the appearance inspection device according to the first aspect, as described above, the processing unit is configured to perform a detection filter process to reduce the detection sensitivity for defective parts as the degree of roughness of the surface of the object shown in the appearance image increases, thereby generating a detection image according to the degree of roughness of the surface of the object shown in the appearance image. In this way, when the degree of roughness of the object is small, the detection sensitivity for defective parts can be fixed to generate a detection image. In addition, when the degree of roughness of the object is large and a part that is not a defective part is likely to be erroneously detected as a defective part, the detection sensitivity for defective parts can be reduced to generate a detection image. As a result, even when the degree of surface roughness varies depending on the object, information for appropriately detecting defective parts in an appearance image showing the object can be obtained.

[0009] In the appearance inspection apparatus according to the first aspect, preferably, the processing unit performs detection filter processing to reduce the threshold value for detecting a defective portion so as to reduce the detection sensitivity for the defective portion included in the appearance image as the roughness of the surface of the object shown in the appearance image increases, thereby generating a detection image corresponding to the roughness of the surface of the object shown in the appearance image. With this configuration, when the roughness of the object is small, the detection image can be generated by fixing the threshold value for detecting the defective portion to a predetermined value. Also, when the roughness of the object is large and it is easy to erroneously detect a non-defective portion as a defective portion, the detection image can be generated by setting the threshold value for detecting the defective portion to be smaller than the predetermined value. As a result, even when the degree of surface roughness varies depending on the object, a detection image for detecting a defective portion generated using a threshold value corresponding to the degree of surface roughness of the object can be obtained.

[0010] In this case, preferably, the processing unit generates a determination image obtained by binarizing the pixel values of the appearance image based on a determination threshold value that is a predetermined pixel value, and is configured to determine the degree of roughness of the surface of the object shown in the appearance image based on the determination image. With this configuration, it is possible to easily determine whether the degree of roughness of the object is large based on the determination image, so that the threshold value for generating the detection image can be easily set according to the degree of roughness of the object. As a result, information for detecting a defective portion of the appearance image showing the object can be easily obtained.

[0011] In the configuration in which the processing unit generates the judgment image, the processing unit is preferably configured to obtain a black dot ratio, which is the ratio of the area of ​​the judgment image occupied by the area of ​​the binarized parts with low pixel values, and to determine that the degree of roughness of the surface of the object shown in the appearance image is small when the black dot ratio is smaller than a predetermined reference ratio, and to determine that the degree of roughness of the surface of the object shown in the appearance image is large when the black dot ratio is equal to or larger than the reference ratio. With this configuration, it is possible to appropriately determine whether the degree of surface roughness of the object is large based on the black dot ratio, which is the ratio of the parts with low pixel values ​​obtained based on the judgment image. As a result, it is possible to appropriately obtain information for detecting defective parts of the appearance image in which the object is shown.

[0012] In this case, the processing unit is preferably configured to generate a first detection image in which pixel values ​​are binarized based on a first pixel threshold value for the appearance image when it is determined that the degree of roughness of the surface of the object shown in the appearance image is small, and to generate a second detection image in which pixel values ​​are binarized based on a second pixel threshold value smaller than the first pixel threshold value for the appearance image when it is determined that the degree of roughness of the surface of the object shown in the appearance image is large. With this configuration, it is possible to obtain a detection image based on a threshold value according to the degree of roughness of the surface of the object. As a result, it is possible to generate an appropriate detection image even when the degree of roughness of the surface of the object varies.

[0013] In the appearance inspection device in which the processing unit generates the first detection image or the second detection image, the processing unit is preferably configured to further generate a third detection image in which pixel values ​​are binarized based on a third pixel threshold value that is greater than the first pixel threshold value. With this configuration, even if the pixel values ​​of a defective portion included in the object are higher than the pixel values ​​of a portion of the object without defects, the detection image can be appropriately generated based on the third pixel threshold value.

[0014] In this case, the processing unit is preferably configured to detect, as defective parts in the appearance image, a portion of pixels in the first detection image that are smaller than the first pixel threshold value, a portion of pixels in the second detection image that are smaller than the second pixel threshold value, and a portion of pixels in the third detection image that are larger than the third pixel threshold value. With this configuration, by detecting, as defective parts in the appearance image, a portion of pixels in the first detection image that are smaller than the first pixel threshold value, and a portion of pixels in the second detection image that are smaller than the second pixel threshold value, defective parts can be easily detected based on black parts with small pixel values ​​in the first detection image and the second detection image. Also, by detecting, as defective parts in the appearance image, a portion of pixels in the third detection image that are larger than the third pixel threshold value, defective parts can be easily detected based on white parts with large pixel values ​​in the third detection image.

[0015] In the appearance inspection device according to the first aspect, preferably, the appearance image includes a plurality of regions, and the processing unit is configured to perform a filter process to reduce the detection sensitivity for the defect portion included in each of the plurality of regions as the degree of roughness of the surface of the object captured in each of the plurality of regions increases, thereby generating a plurality of detection images corresponding to each of the plurality of regions according to the degree of roughness of the surface of the object captured in each of the plurality of regions. With this configuration, even if there are parts in the appearance image where the degree of roughness of the object captured in the appearance image differs, an appropriate detection image can be generated for each region according to the degree of roughness of each region that divides the appearance image.

[0016] In the appearance inspection device according to the first aspect, the processing unit is preferably configured to detect the position and shape of the defect in the appearance image based on the detection image. With this configuration, even if the degree of roughness of the surface of the object is not constant, the defect in the appearance image in which the object is shown can be properly detected. As a result, the position and shape of the defect in the appearance image can be properly confirmed.

[0017] In the appearance inspection device in which the processing unit detects the position and shape of a defect in the appearance image, the processing unit is preferably configured to perform a size filter process on the detection image so as to ignore defects smaller than a predetermined size, thereby generating a size-filtered image, and to detect the position and shape of the defect in the appearance image based on the size-filtered image. With this configuration, it is possible to suppress excessive detection of defects smaller than a predetermined size that do not affect the performance of the product, or parts that are erroneously detected as defects due to noise in the appearance image, for example. As a result, it is possible to more appropriately detect defects in the appearance image.

[0018] The appearance inspection method according to the second aspect includes an image acquisition step of acquiring an appearance image of the object, and a detection image generation step of generating a detection image corresponding to the degree of surface roughness of the object shown in the appearance image by performing filter processing such that the detection sensitivity for defects contained in the appearance image decreases as the degree of surface roughness of the object shown in the appearance image increases.

[0019] The appearance inspection method according to the second aspect includes a detection image generating step of generating a detection image according to the degree of surface roughness of the object shown in the appearance image by performing a filter process to reduce the detection sensitivity for the defective part included in the appearance image as the degree of roughness of the surface of the object shown in the appearance image increases, as described above. In this way, when the degree of roughness of the object is small, the detection image can be generated with a fixed detection sensitivity for the defective part. In addition, when the degree of roughness of the object is large and a part that is not a defective part is likely to be erroneously detected as a defective part, the detection sensitivity for the defective part can be reduced to generate the detection image. As a result, it is possible to provide an appearance inspection method capable of acquiring information for appropriately detecting a defective part in an appearance image showing an object, even when the degree of surface roughness varies depending on the object. Effect of the Invention

[0020] According to the present invention, as described above, it is possible to provide an appearance inspection device and an appearance inspection method that are capable of acquiring information for appropriately detecting defective areas in an appearance image depicting an object, even when the degree of surface roughness varies depending on the object. [Brief description of the drawings]

[0021] [Figure 1] FIG. 1 is a schematic diagram showing a visual inspection apparatus according to a first embodiment. [Diagram 2] 4A to 4C are diagrams for explaining appearance images when the degree of roughness of the surface of the semiconductor wafer is small according to the first embodiment. [Diagram 3] 4A to 4C are diagrams for explaining appearance images of a semiconductor wafer having a highly rough surface according to the first embodiment. [Figure 4] 4 is a flowchart for explaining a process of a visual inspection method according to the first embodiment. [Diagram 5] 4A to 4C are diagrams for explaining an appearance image and a defective portion when the degree of roughness of the surface of the semiconductor wafer is small according to the first embodiment. [Figure 6] FIG. 4 is a diagram for explaining a high-pass image in a case where the degree of roughness of the surface of the semiconductor wafer is small according to the first embodiment. [Figure 7] 5A to 5C are diagrams for explaining determination images in a case where the degree of roughness of the surface of the semiconductor wafer is small according to the first embodiment. [Figure 8] 4 is a diagram for explaining a first detection image when the degree of roughness of the surface of the semiconductor wafer is small according to the first embodiment. FIG. [Figure 9] 10 is a diagram for explaining a first black spot detection image when the degree of roughness of the surface of the semiconductor wafer is small according to the first embodiment. FIG. [Figure 10] 13 is a diagram for explaining a third detection image in a case where the degree of roughness of the surface of the semiconductor wafer is small according to the first embodiment. FIG. [Figure 11]4A to 4C are diagrams for explaining appearance images of a semiconductor wafer having a highly rough surface according to the first embodiment. [Figure 12] 5A to 5C are diagrams for explaining determination images when the surface of a semiconductor wafer is highly rough in accordance with the first embodiment. [Figure 13] 6 is a diagram for explaining a second detection image in a case where the surface of the semiconductor wafer is largely rough according to the first embodiment. FIG. [Figure 14] 13A and 13B are diagrams for explaining appearance images in which the surface of a semiconductor wafer has both large and small roughness levels according to the second embodiment. [Figure 15] FIG. 11 is a diagram for explaining division of an appearance image into a plurality of regions according to the second embodiment. [Figure 16] FIG. 11 is a diagram for explaining an example of an area where the surface roughness of a semiconductor wafer is small according to the second embodiment. [Figure 17] 10 is a flowchart for explaining a process of a visual inspection method according to a second embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0022] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings.

[0023] [First embodiment] The configuration of a visual inspection apparatus 100 according to the first embodiment will be described with reference to FIG.

[0024] (Configuration of visual inspection device) 1, the appearance inspection apparatus 100 includes a processing unit 1, an image acquisition unit 5, and a display unit 6. The processing unit 1 includes a determination unit 2, a filter processing unit 3, and a detection unit 4 as functional blocks configured by software. In the first embodiment, the appearance inspection apparatus 100 is configured to inspect the appearance of a surface S of a semiconductor wafer U.

[0025] The semiconductor wafer U is, for example, a silicon wafer obtained by thinning single crystal silicon (Si). In the first embodiment, the surface S of the semiconductor wafer U is plated, and gold (Au) plating is formed. In the first embodiment, the surface S of the semiconductor wafer U may have defects D including not only partial defects and scratches, but also dirt and foreign matter adhesion. The surface S of the semiconductor wafer U varies in roughness due to the density of the gold plating and the size of the grain boundaries. Here, "roughness" refers to a portion where the pixel value is not constant and appears to vary in the captured image, even if there is no defect D on the surface S of the semiconductor wafer U. The semiconductor wafer U is an example of an "object" in the claims.

[0026] 1 includes a processor such as a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and a GPU (Graphics Processing Unit). The processing unit 1 is configured to read an appearance image 10 or an appearance image 20 (see FIGS. 2 and 3) acquired by an image acquisition unit 5 described later, and to perform image processing on the appearance image 10 or the appearance image 20, thereby performing an appearance inspection of a surface S of a semiconductor wafer U.

[0027] 1, the processing unit 1 includes a determination unit 2. The determination unit 2 is configured to determine the degree of roughness on the surface S of the semiconductor wafer U based on the acquired determination image 10b (see FIG. 7) or determination image 20b (see FIG. 12). The detailed operation of the determination unit 2 will be described later.

[0028] As shown in Fig. 1, the processing unit 1 includes a filter processing unit 3. The filter processing unit 3 is configured to generate information (images) for detecting a defect portion 12 or a defect portion 22 (see Figs. 2 and 3) which is a portion in which a defect D is captured in the appearance image 10 or the appearance image 20, based on the degree of roughness in the surface S of the semiconductor wafer U captured in the appearance image 10 or the appearance image 20 (see Figs. 2 and 3). The filter processing unit 3 is also configured to perform various types of filter processing, such as a high-pass filter or a binarization filter, on the appearance image 10 or the appearance image 20. The detailed operation of the filter processing unit 3 will be described later.

[0029] 1, the processing unit 1 includes a detection unit 4. The detection unit 4 is configured to detect a defect portion 12 and a defect portion 22 (see FIGS. 2 and 3) in the appearance image 10 or the appearance image 20 based on the image generated by the filter processing unit 3. The detailed operation of the detection unit 4 will be described later.

[0030] The image acquisition unit 5 is configured to photograph the surface S of the semiconductor wafer U and acquire the appearance image 10 or the appearance image 20. In the first embodiment, the image acquisition unit 5 includes a camera, and acquires the appearance of the surface S of the semiconductor wafer U photographed by the camera as the appearance image 10 or the appearance image 20 expressed in grayscale. In the first embodiment, the image acquisition unit 5 is connected to the processing unit 1 by a communication line, and transmits the acquired appearance image 10 or the appearance image 20 to the processing unit 1.

[0031] In the first embodiment, the display unit 6 is configured to display the appearance image 10 or the appearance image 20 acquired by the image acquisition unit 5, and various images generated by applying filter processing to the appearance image 10 or the appearance image 20 by the processing unit 1. The display unit 6 is, for example, a display device such as a liquid crystal monitor.

[0032] (Exterior image) Here, referring to FIG. 2 and FIG. 3, the appearance image 10 or the appearance image 20 showing the surface S of the semiconductor wafer U will be described. For example, as shown in FIG. 2, when the degree of roughness of the surface S of the semiconductor wafer U is small, the difference in pixel value between the normal part 11 showing the part where the defect D is not shown in the appearance image 10 and the defective part 12 showing the part where the defect D is shown in the appearance image 10 is large. On the other hand, as shown in FIG. 3, when the degree of roughness of the surface S of the semiconductor wafer U is large, the difference in pixel value between the normal part 21 showing the part where the defect D is not shown in the appearance image 20 and the defective part 22 showing the defect D in the appearance image 20 is small. In this case, since the difference between the normal part 21 and the defective part 22 is unclear, the defective part 22 may not be detected appropriately. In addition, the normal part 21 in the appearance image 20 may be erroneously detected as the defective part 22.

[0033] (Appearance inspection method) Next, a description will be given of a method for visually inspecting the surface S of a semiconductor wafer U in the first embodiment. A processing unit 1 of the visual inspection apparatus 100 shown in Fig. 1 executes the process of the visual inspection method for the surface S of a semiconductor wafer U shown in Fig. 4.

[0034] (Appearance inspection method when the degree of roughness of the surface S of a semiconductor wafer U is small) First, a case where the surface S of a semiconductor wafer U is only slightly rough and includes a defect 12 with small pixels and a defect 13 with large pixels in an appearance image 10 will be described with reference to FIG. 1 and FIGS.

[0035] First, as an image acquisition step of step S1, the image acquisition unit 5 shown in Fig. 1 images the surface S of the semiconductor wafer U. In the first embodiment, the semiconductor wafer U is fixed to a stage (not shown), and the image acquisition unit 5 is installed vertically above the surface S of the semiconductor wafer U. Then, the image acquisition unit 5 images the surface S of the semiconductor wafer U from above to acquire an appearance image 10 shown in Fig. 5. Then, the processing unit 1 acquires the appearance image 10 captured by the image acquisition unit 5. After that, the process proceeds to step S2.

[0036] Next, as a high-pass filter processing step of step S2, the filter processing unit 3 shown in Fig. 1 performs high-pass filter processing on the acquired appearance image 10. Through this high-pass filter processing, the filter processing unit 3 cuts low-frequency components in the appearance image 10 to accentuate edges (boundary portions in the appearance image 10 where pixel values ​​change significantly) and generates a high-pass image 10a as shown in Fig. 6. Thereafter, the process proceeds to step S3.

[0037] Next, as a judgment image generating step of step S3, the filter processing unit 3 performs a binarization filter process on the high-pass image 10a. In this first embodiment, the pixel value (judgment threshold value) used in the binarization filter process is fixed to a certain value, and is a pixel value of 100 among the pixel values ​​of the grayscale expressed in 256 gradations. That is, in the high-pass image 10a, a portion having a pixel value of 100 or more is represented as white (pixel value of 255), and a pixel value less than 100 is represented as black (pixel value of 0). By performing this binarization filter process, the filter processing unit 3 generates a judgment image 10b as shown in FIG. 7. Then, the process proceeds to step S4.

[0038] Next, as a judgment step of step S4, the judgment unit 2 judges the degree of roughness of the surface S of the semiconductor wafer U based on the generated judgment image 10b. The judgment unit 2 judges the degree of roughness of the surface S of the semiconductor wafer U based on the black dot ratio, which is the ratio of the area of ​​the black dots 14 with a pixel value of 0 to the image area of ​​the judgment image 10b. In this first embodiment, the judgment unit 2 judges that the "degree of roughness is small" when the black dot ratio of the judgment image 10b is less than the reference ratio of 30%. Also, the judgment unit 2 is configured to judge that the "degree of roughness is large" when the black dot ratio of the judgment image 10b is equal to or greater than the reference ratio of 30%. In this first embodiment, the judgment image 10b in FIG. 7 has a black dot ratio of 15%, and the judgment unit 2 judges that the "degree of roughness is small" for the surface S of the semiconductor wafer U, and proceeds to the processing of step S5a in the flow of FIG. 4.

[0039] Next, as a detection image generating step of step S5a, the filter processing unit 3 performs a detection filter process in which the detection sensitivity for the defect portion 12 included in the appearance image 10 is changed based on the judgment result of the judgment unit 2. Specifically, the filter processing unit 3 is configured to perform a binarization filter process for the appearance image 10, which reduces the threshold value of the pixel value so that the detection sensitivity for the defect portion 12 decreases as the degree of roughness of the surface S of the semiconductor wafer U increases. In the first embodiment, when the judgment unit 2 judges that the "degree of roughness is small", the filter processing unit 3 performs a binarization process for the appearance image 10 using the pixel value 120 as the first pixel threshold value. When the judgment unit 2 judges that the "degree of roughness is large", the filter processing unit 3 performs a binarization process for the appearance image 10 using the pixel value 80 as the second pixel threshold value. In this first embodiment, since the filter processing unit 3 has obtained the determination result that "the degree of roughness is small," the filter processing unit 3 performs binarization filter processing using the first pixel threshold value on the appearance image 10 to generate the first detection image 10c shown in Fig. 8. Then, the process proceeds to step S6.

[0040] Next, as a size filter processing step of step S6, the filter processing unit 3 performs size filter processing on the generated first detection image 10c so as not to detect black spots 14 having an area smaller than a predetermined area as defective parts 12. In this first embodiment, the filter processing unit 3 performs size filter processing on the first detection image 10c to generate a first black spot detection image 10d from which all black spots 14 other than defective parts 12 have been removed. Note that the first black spot detection image 10d is an example of the "first detection image" in the claims. Then, the process proceeds to step S7.

[0041] Next, as a detection step of step S7, the detection unit 4 shown in FIG. 1 detects the defective portion 12 in the appearance image 10 based on the generated first black spot detection image 10d. In this first embodiment, the detection unit 4 detects a portion (black portion) in the first black spot detection image 10d where the pixel value is 0 as the defective portion 12, and detects the position and shape of the defective portion 12 by corresponding it to the position of the appearance image 10. Also, in this first embodiment, the detection unit 4 is configured to display the position and shape of the defective portion 12 on the display unit 6 and to notify by sound that the defective portion 12 has been found. This allows an operator or the like to grasp the presence of the defect D on the surface S of the semiconductor wafer U. Then, the process proceeds to step S8.

[0042] Next, in the detection image generating step of step S8, binarization filter processing is performed on the appearance image 10 using a third pixel threshold value having a pixel value greater than the first pixel threshold value to detect the defective portion 13. In this first embodiment, the third pixel threshold value is 180, and a third detection image 10e is generated in which parts of the appearance image 10 with a pixel value less than 180 are displayed as black (pixel value 0) and parts with a pixel value equal to or greater than 180 are displayed as white dots (pixel value 0). Then, the process proceeds to step S9.

[0043] Next, as a size filter processing step of step S9, the filter processing unit 3 performs size filter processing on the generated third detection image 10e so as not to detect white dots smaller than a predetermined area as defective parts 13. In this first embodiment, the filter processing unit 3 performs size filter processing on the third detection image 10e so that the image generated after the size filter processing is the same as the third detection image 10e, since there are no white dots smaller than a predetermined area in the third detection image 10e. Then, the process proceeds to step S10.

[0044] Next, as a detection step of step S10, the detection unit 4 detects the defect portion 13 in the appearance image 10 based on the generated third detection image 10e. In this first embodiment, the detection unit 4 detects a portion (white portion) in the third detection image 10e with a pixel value of 255 as the defect portion 13, and identifies the position and shape of the defect D by corresponding it to the position of the appearance image 10. Furthermore, in the detection step of step S7, the detection unit 4 is configured to display the position and shape of the defect portion 13 detected in the detection step of step S10 together with the defect portion 12 displayed on the display unit 6, and to notify again by sound that the defect portion 13 has been found. Thereafter, when performing an appearance inspection of another position on the surface S of the semiconductor wafer U, the relative positions of the image acquisition unit 5 and the semiconductor wafer U are changed by a stage or the like (not shown), and the operations from step S1 are repeated.

[0045] (Appearance inspection method when the surface S of a semiconductor wafer U is highly rough) Next, a case will be described where the surface S of the semiconductor wafer U is highly rough and no defect 22 (see FIG. 3) is captured in the appearance image 20, as in the appearance image 20 shown in FIG 11. Explanations of steps common to the above-mentioned case where the surface S of the semiconductor wafer U is only slightly rough and there are small pixel defect 12 and large pixel defect 13 will be omitted.

[0046] Even when the degree of roughness of the surface S of the semiconductor wafer U is large and no defective portion 22 exists, in step S4 shown in FIG. 4, the judgment unit 2 judges the degree of roughness of the surface S of the semiconductor wafer U based on the generated judgment image 20b shown in FIG. 12. The judgment unit 2 is configured to judge that the degree of roughness is small when the black dot ratio, which is the ratio of the area of ​​the parts with pixel values ​​of 0 in the image area of ​​the judgment image 20b, is less than the reference ratio of 30%, and to judge that the degree of roughness is large when the black dot ratio is equal to or greater than the reference ratio of 30%. Here, the judgment image 20b in FIG. 12 has a black dot ratio of 40%. Therefore, the judgment unit 2 judges that the degree of roughness is large for the surface S of the semiconductor wafer U, and proceeds to the processing of step S5b in the flow of FIG. 4.

[0047] Next, as a detection image generating step of step S5b, the filter processing unit 3 performs a detection filter process in which the detection sensitivity for a non-existent defect portion 22 is changed based on the judgment result of the judgment unit 2. In this case, the filter processing unit 3 performs a binarization filter process on the appearance image 20 using a pixel value 80 as a second pixel threshold value as a threshold value based on the judgment of the judgment unit 2 that the degree of roughness is large. This generates a second detection image 20c shown in FIG. 13. The subsequent process is the same as the above-mentioned explanation of the case where the degree of roughness of the surface S of the semiconductor wafer U is small and there is a defect portion 12 with small pixels and a defect portion 13 with large pixels. Note that since the defect portion 22 does not exist in the appearance image 20, the detection unit 4 does not detect the defect portion 22 and ends the appearance inspection. In this case, the detection unit 4 causes the display unit 6 to display that all the parts captured in the appearance image 20 were normal parts 21.

[0048] In the manner described above, the visual inspection apparatus 100 determines the degree of roughness of the surface S of the semiconductor wafer U, whether the degree of roughness is large or small, and performs binarization filter processing with different threshold values ​​based on the determination result, thereby generating detection image 10c or detection image 20c as information for detecting defective portion 12, defective portion 13, and defective portion 22.

[0049] (Effects of the first embodiment) In the first embodiment, the following effects can be obtained.

[0050] The appearance inspection device 100 of the first embodiment includes an image acquisition unit 5 that acquires an appearance image 10 or an appearance image 20 of a semiconductor wafer U, and a processing unit 1 that performs a detection filter process on the appearance image 10 or the appearance image 20 to generate a detection image 10c for detecting a defect portion 12 and a defect portion 13 or a detection image 20c for detecting a defect portion 22, and the processing unit 1 is configured to perform a detection filter process such that the detection sensitivity for the defect portion 12 or the defect portion 22 decreases as the degree of roughness of the surface S of the semiconductor wafer U captured in the appearance image 10 or the appearance image 20 increases, thereby generating a detection image 10c or a detection image 20c according to the degree of roughness of the surface S of the semiconductor wafer U captured in the appearance image 10 or the appearance image 20. As a result, when the degree of roughness of the semiconductor wafer U is small, the detection sensitivity for the defect portion 12 can be fixed and the detection image 10c can be generated. Furthermore, when the degree of roughness of the semiconductor wafer U is large and normal portion 21 is likely to be erroneously detected as defective portion 22, the detection sensitivity for defective portion 22 can be lowered or detection image 20c can be generated. As a result, even when the degree of roughness of surface S varies depending on the semiconductor wafer U, information for appropriately detecting defective portion 12 and defective portion 22 in appearance image 10 and appearance image 20 in which the object is captured can be acquired.

[0051] In the first embodiment, the processing unit 1 performs detection filter processing to reduce the threshold value for detecting the defective portion 12 or the defective portion 22 so as to reduce the detection sensitivity for the defective portion 12 included in the appearance image 10 or the defective portion 22 included in the appearance image 20 as the roughness of the surface S of the semiconductor wafer U shown in the appearance image 10 or the appearance image 20 increases. As a result, a detection image 10c or a detection image 20c corresponding to the roughness of the surface S of the semiconductor wafer U shown in the appearance image 10 or the appearance image 20 is generated. Thereby, when the roughness of the semiconductor wafer U is small, the detection image 10c can be generated by fixing the threshold value for detecting the defective portion 12 to a predetermined value. Further, when the roughness of the semiconductor wafer U is large and the normal portion 21 is likely to be erroneously detected as the defective portion 22, the detection image 20c can be generated by setting the threshold value for detecting the defective portion 22 to be smaller than a predetermined value. As a result, even when the degree of surface roughness varies depending on the semiconductor wafer U, a detection image 10c or a detection image 20c for detecting the defective portions 12, 13, and 22 can be generated using a threshold value corresponding to the roughness of the surface S of the semiconductor wafer U.

[0052] Also, in the first embodiment, the processing unit 1 generates a determination image 10b or a determination image 20b obtained by binarizing the pixel values of the appearance image 10 or the appearance image 20 based on a determination threshold value that is a predetermined pixel value, and determines the roughness of the surface S of the semiconductor wafer U shown in the appearance image 10 or the appearance image 20 based on the determination image 10b or the determination image 20b. Thereby, it is possible to easily determine whether the roughness of the semiconductor wafer U is large based on the determination image 10b or the determination image 20b, so that the threshold value for generating the detection image 10c or the detection image 20c can be easily set according to the roughness of the semiconductor wafer U. As a result, information for detecting the defective portions 12, 13, and 22 of the appearance image 10 or the appearance image 20 in which the semiconductor wafer U is shown can be easily obtained.

[0053] In the first embodiment, the processing unit 1 is configured to acquire a black dot ratio, which is the ratio of the area of ​​the determination image 10b or the determination image 20b occupied by the area of ​​the binarized low pixel value portion, and determine that the degree of roughness of the surface S of the semiconductor wafer U shown in the appearance image 10 or the appearance image 20 is small when the black dot ratio is smaller than a predetermined reference ratio, and determine that the degree of roughness of the surface S of the semiconductor wafer U shown in the appearance image 10 or the appearance image 20 is large when the black dot ratio is equal to or larger than the reference ratio. This makes it possible to appropriately determine whether the degree of roughness of the surface S of the semiconductor wafer U is large based on the black dot ratio, which is the ratio of the area of ​​the low pixel value portion obtained based on the determination image 10b or the determination image 20b. As a result, it is possible to appropriately acquire information for detecting the defective portion 12 and the defective portion 22 of the appearance image 10 or the appearance image 20 in which the semiconductor wafer U is shown.

[0054] In the first embodiment, the processing unit 1 is configured to generate a first detection image 10c in which the pixel values ​​are binarized based on a first pixel threshold value for the appearance image 10 or the appearance image 20 when it is determined that the degree of roughness of the surface S of the semiconductor wafer U shown in the appearance image 10 or the appearance image 20 is small, and to generate a second detection image 20c in which the pixel values ​​are binarized based on a second pixel threshold value smaller than the first pixel threshold value for the appearance image 10 or the appearance image 20 when it is determined that the degree of roughness of the surface S of the semiconductor wafer U shown in the appearance image 10 or the appearance image 20 is large. This makes it possible to obtain the detection image 10c or the detection image 20c based on a threshold value according to the degree of roughness of the surface S of the semiconductor wafer U. As a result, even when the degree of roughness of the surface S of the semiconductor wafer U is different, it is possible to generate an appropriate detection image 10c or detection image 20c.

[0055] In the first embodiment, the processing unit 1 is configured to further generate a third detection image 10e in which pixel values ​​are binarized based on a third pixel threshold value that is greater than the first pixel threshold value. This makes it possible to appropriately generate the detection image 10e based on the third pixel threshold value even when the pixel value of a defective portion 13 included in the semiconductor wafer U is greater than the pixel value of a normal portion 11 in the semiconductor wafer U.

[0056] In the first embodiment, the processing unit 1 is configured to detect the pixel portion of the first detection image 10c having a pixel value smaller than the first pixel threshold value, the pixel portion of the second detection image 20c having a pixel value smaller than the second pixel threshold value, and the pixel portion of the third detection image 10e having a pixel value larger than the third pixel threshold value as the defect portion 12, the defect portion 13, or the defect portion 22 in the appearance image 10 or the appearance image 20. With this configuration, by detecting the pixel portion of the first detection image 10c having a pixel value smaller than the first pixel threshold value and the pixel portion of the second detection image 20c having a pixel value smaller than the second pixel threshold value as the defect portion 12 or the defect portion 22 in the appearance image 10 or the appearance image 20, the defect portion 12 or the defect portion 22 can be easily detected based on the black portions with small pixel values ​​in the first detection image 10c and the second detection image 20c. Furthermore, if the portion of the third detection image 10e whose pixels are larger than the third pixel threshold value is configured to be detected as a defective portion 13 in the appearance image 10 or the appearance image 20, the defective portion 13 can be easily detected based on the white portion with a large pixel value in the third detection image 10e.

[0057] Moreover, in the first embodiment, the processing unit 1 is configured to detect the position and shape of the defect portion 12 and the defect portion 13 in the appearance image 10, or the defect portion 22 in the appearance image 20, based on the detection image 10c, the detection image 20c, and the detection image 10e. This makes it possible to properly detect the defect portion 12 and the defect portion 13 in the appearance image 10 in which the semiconductor wafer U is shown, or the defect portion 22 in the appearance image 20, even if the degree of roughness of the surface S of the semiconductor wafer U is not constant. As a result, it is possible to properly confirm the position and shape of the defect portion 12, the defect portion 13, and the defect portion 22 in the appearance image 10 or the appearance image 20.

[0058] In the first embodiment, the processing unit 1 is configured to perform a size filter process on the detection image 10c and the detection image 20c so as to ignore black dots 14 smaller than a predetermined size, thereby generating a size-filtered image, and detect the position and shape of the defective portion 12 in the appearance image 10 or the defective portion 22 in the appearance image 20 based on the size-filtered image. This makes it possible to suppress excessive detection of black dots 14 smaller than a predetermined size that does not affect the performance of the product, or black dots 14 that are erroneously detected as defective portions 12 due to noise in the appearance image 10 or the appearance image 20. As a result, the defective portions 12 and 22 in the appearance image 10 or the appearance image 20 can be more appropriately detected.

[0059] [Second embodiment] Next, the appearance inspection apparatus 100a and the appearance inspection method according to the second embodiment will be described. The apparatus configuration of the appearance inspection apparatus 100a used in the appearance inspection method according to the second embodiment is the same as the appearance inspection apparatus 100 shown in FIG. 1, except for the processing unit 1a. In the second embodiment, a case will be described in which the degree of roughness varies within the surface of an appearance image 30 (see FIG. 14) of the acquired surface S of the semiconductor wafer U. Specifically, as shown in FIG. 14, a case will be described in which a defective portion 32 in which the defect D shown in FIG. 1 is captured is detected from an appearance image 30 in which the degree of roughness of the surface S of the semiconductor wafer U is small and a normal portion 31 in which the defect D shown in FIG. 1 is not captured and a normal portion 41 in which the degree of roughness of the surface S of the semiconductor wafer U is large and a defect D shown in FIG. 1 is not captured are mixed. In this second embodiment, steps S1a and S1b are added between steps S1 and S2, as in the processing of the appearance inspection method of the second embodiment shown in FIG. 17. Note that, in the second embodiment, a description of points common to the first embodiment will be omitted.

[0060] In the second embodiment, first, as step S1 in Fig. 17, the processing unit 1a acquires the appearance image 30, and then, as an image division step of step S1a, divides the appearance image 30 into a plurality of regions. Specifically, in the second embodiment, the processing unit 1a divides the appearance image 30 into 5 x 5 rectangular regions as shown in Fig. 15. Note that the fineness with which the appearance image 30 is divided can be changed as desired.

[0061] Next, as a determination data acquisition step of step S1b, the processing unit 1a is configured to acquire determination data for each of the divided regions of the appearance image 30 to determine whether the degree of roughness varies within the surface of the appearance image 30. Specifically, in this second embodiment, the processing unit 1a acquires average pixel value data for each divided region, and checks whether the average pixel value data in each divided region falls within a predetermined amount of variation. Then, when the average pixel value data in all regions falls within the predetermined amount of variation, the processing unit 1a judges that "the degree of roughness does not vary within the appearance image 30". Furthermore, when there is a mixture of a region in which the average pixel value data does not fall within the predetermined amount of variation and a region in which the average pixel value data falls within the predetermined amount of variation, the processing unit 1a judges that "the degree of roughness varies within the appearance image 30". For example, in this embodiment, it is checked whether the average pixel value falls within a range of 100±20. 15, the average pixel value of divided area 31a is 90, and the average pixel value of divided area 41a is 70. Therefore, since there are areas in which the average pixel value data does not fall within a predetermined variation amount and areas in which the average pixel value data falls within a predetermined variation amount, processing unit 1a determines that "the degree of roughness varies within the appearance image 30."

[0062] In the second embodiment, since the degree of roughness varies within the surface of the appearance image 30, as shown in FIG. 17, the same processing as steps S2 to S10 described in the first embodiment is performed for one divided region. Furthermore, after the inspection is completed by the processing of steps S2 to S10 for one divided region, the process proceeds to step 11. In step 11, it is determined whether the processing of steps S2 to S10 has been performed for all divided regions. If there is an unprocessed region, the processing of steps S2 to S10 is performed again for the unprocessed region until the processing is completed for all regions. Also, for example, as shown in FIG. 16, when a defect 32 is detected in the region 31a, the display unit 6 displays the position of the defect 32 in the appearance image 30 together with the image of the region 31a. By repeating this for each divided region, even if the degree of roughness varies within the surface of the appearance image 30 (see FIG. 14) of the surface S of the semiconductor wafer U, an appearance inspection according to the degree of roughness is performed.

[0063] (Effects of the second embodiment) Next, the effects of the second embodiment will be described.

[0064] In the second embodiment, the appearance image 30 includes a plurality of regions, and the processing unit 1a is configured to perform a filter process to reduce the detection sensitivity for the defect portion 32 included in each of the plurality of regions as the degree of roughness of the surface S of the semiconductor wafer U reflected in each of the plurality of regions increases, thereby generating a plurality of detection images corresponding to each of the plurality of regions according to the degree of roughness of the surface S of the semiconductor wafer U reflected in each of the plurality of regions. As a result, even if there are parts in the appearance image 30 where the degree of roughness of the semiconductor wafer U reflected in the appearance image 30 differs, an appropriate detection image can be generated for each region according to the degree of roughness of each region separating the appearance image 30.

[0065] Other effects of the second embodiment are similar to those of the first embodiment.

[0066] [Variations] The embodiments disclosed herein should be considered to be illustrative and not restrictive in all respects. The scope of the present invention is indicated by the claims, not by the description of the embodiments above, and further includes all modifications (variations) within the meaning and scope of the claims.

[0067] For example, in the above first and second embodiments, the appearance inspection apparatus 100 and the appearance inspection apparatus 100a perform an appearance inspection of the plated surface S of the semiconductor wafer U, but the present invention is not limited to this. In the present invention, the object to be subjected to the appearance inspection may be anything, for example, it may be the unplated surface of the semiconductor wafer U, or it may be just a metal material different from the semiconductor wafer U.

[0068] In the above first and second embodiments, the appearance inspection apparatus 100 and the appearance inspection apparatus 100a use the image acquisition unit 5 including a camera to capture an image of the surface S of the semiconductor wafer U, but the present invention is not limited to this. In the present invention, the image acquisition unit 5 may be configured to read and capture data of the appearance images 10 to 30 captured by another device.

[0069] In the above first and second embodiments, an example in which the display unit 6 is provided has been shown, but the present invention is not limited to this. In the present invention, the display unit 6 may be provided in a separate device, or may not be provided anywhere.

[0070] In addition, in the above first and second embodiments, an example was shown in which the processing unit 1 or processing unit 1a performs detection filter processing to reduce the threshold value for detecting the defect portion 12, defect portion 22 or defect portion 32 contained in the appearance images 10 to 30 as the degree of roughness of the surface S of the semiconductor wafer U captured in the appearance images 10 to 30 increases, thereby generating a detection image 10c or a detection image 20c corresponding to the degree of roughness of the surface S of the semiconductor wafer U captured in the appearance images 10 to 30, but the present invention is not limited to this. In the present invention, for example, when pixel values ​​appear high in portions of the surface S of the semiconductor wafer U in the appearance images 10 to 30 where the degree of roughness is large, the processing unit 1 or processing unit 1a may be configured to perform detection filter processing that increases the threshold value for detecting large defective portions 13 of the pixels included in the appearance images 10 to 30 as the degree of roughness of the surface S of the semiconductor wafer U appearing in the appearance images 10 to 30 increases, thereby generating a detection image corresponding to the degree of roughness of the surface S of the semiconductor wafer U appearing in the appearance images 10 to 30, or a combination of these may be used.

[0071] In the above first and second embodiments, the processing unit 1 or the processing unit 1a generates the judgment image 10b or the judgment image 20b by binarizing the pixel values ​​of the appearance images 10 to 30 based on the pixel value of 100, which is a judgment threshold value that is a predetermined pixel value, and judges the degree of roughness of the surface S of the semiconductor wafer U shown in the appearance images 10 to 30 based on the judgment image 10b or the judgment image 20b, but the present invention is not limited to this. In the present invention, the judgment threshold value may be set to an arbitrary pixel value. In addition, the processing unit 1 or the processing unit 1a may perform any process as long as it can judge the degree of roughness of the surface S of the semiconductor wafer U. For example, the processing unit 1 or the processing unit 1a may be configured to judge using the luminance variance of the appearance images 10 to 30, or to generate a histogram for the pixel values ​​of the appearance images 10 to 30 and judge the degree of roughness of the surface S of the semiconductor wafer U shown in the appearance images 10 to 30 based on the frequency of the histogram.

[0072] In the above first and second embodiments, the detection image 10c or 20c is generated based on the first threshold value when the black spot ratio in the judgment image 10b or the judgment image 20b is equal to or greater than the reference ratio, and based on the second threshold value when the black spot ratio is less than the reference ratio. However, the present invention is not limited to this. In the present invention, two arbitrary first and second reference ratios may be set as the reference ratio of the black spot ratio. In this case, the processing unit 1 may be configured to generate three detection images based on the first threshold value when the black spot ratio in the judgment image 10b or the judgment image 20b is equal to or greater than the first reference ratio, based on the second threshold value when the black spot ratio is equal to or greater than the second reference ratio and less than the first reference ratio, and based on the third threshold value when the black spot ratio is less than the second reference ratio. In addition, three or more reference ratios of the black spot ratio may be set, a threshold value corresponding to each of the reference ratios is set, and a plurality of detection images may be generated based on each of the threshold values.

[0073] In the above first and second embodiments, the processing unit 1 or 1a acquires a black dot ratio, which is the ratio of the area of ​​the determination image 10b or the determination image 20b that is occupied by the area of ​​the parts with low binarized pixel values, and judges the degree of roughness based on the black dot ratio and a predetermined reference ratio, but the present invention is not limited to this. In the present invention, for example, when the pixel values ​​of the parts with high degree of roughness of the surface S of the semiconductor wafer U in the appearance images 10 to 30 are particularly high, a white dot ratio, which is the ratio of the area of ​​the determination image 10b or the determination image 20b that is occupied by the area of ​​the parts with high binarized pixel values, may be acquired, and the degree of roughness may be judged based on the white dot ratio and a predetermined reference ratio.

[0074] In the above-described first and second embodiments, when the processing unit 1 or the processing unit 1a determines that the degree of roughness of the surface S of the semiconductor wafer U shown in the appearance image 10 is small, for the appearance image 10, based on the pixel value 120 which is the first pixel threshold, a first detection image 10c with pixel values binarized is generated. When it is determined that the degree of roughness of the surface S of the semiconductor wafer U shown in the appearance image 20 is large, for the appearance image 20, a second detection image 20c with pixel values binarized based on the pixel value 80 which is the second pixel threshold is generated. However, the present invention is not limited to this. In the present invention, the pixel values of the first pixel threshold and the second pixel threshold may be set arbitrarily.

[0075] Also, in the above-described first and second embodiments, an example is shown in which the processing unit 1 or the processing unit 1a is further configured to generate a third detection image 10e with pixel values binarized based on a pixel value 180 which is a third pixel threshold larger than the first pixel threshold. However, the present invention is not limited to this. In the present invention, the third pixel threshold may be set arbitrarily, or the processing unit 1 or the processing unit 1a may be configured to end the appearance inspection without generating the third detection image 10e.

[0076] Also, in the above-described first and second embodiments, an example is shown in which the processing unit 1 or the processing unit 1a is configured to detect the positions and shapes of the defective portions 12, 13, 22, and 32 in the appearance images 10 to 30 based on the first to third detection images 10c to 30c. However, the present invention is not limited to this. In the present invention, the processing unit 1 or the processing unit 1a may not detect the defective portions 12, 13, 22, and 32, and an operator may detect the positions and shapes of the defective portions 12, 13, 22, and 32 based on the first to third detection images 10c to 30c.

[0077] In the above second embodiment, the appearance image 30 is divided into a plurality of 5×5 rectangular regions, but the present invention is not limited to this. In the present invention, the number of the divided regions may be set arbitrarily. The shapes of the regions may also be set arbitrarily, for example, triangular or trapezoidal.

[0078] In the above first and second embodiments, the processing unit 1 or the processing unit 1a is configured to perform size filtering on the detection image 10c, but the present invention is not limited to this. In the present invention, the processing unit 1 or the processing unit 1a may be configured to perform appearance inspection without performing size filtering.

[0079] In the above first and second embodiments, the processing unit 1 or the processing unit 1a performs high-pass filtering on the appearance images 10 to 30 in step S2 to generate the high-pass image 10a, etc., but the present invention is not limited to this. In the present invention, the processing unit 1 or the processing unit 1a may be configured to perform appearance inspection without performing the high-pass filtering in step S2.

[0080] In the second embodiment, the processor 1 or processor 1a divides the appearance image 30 and determines whether or not there is variation in the degree of roughness of the surface S of the semiconductor wafer U in the appearance image 30 based on the average value of the pixel values ​​of each divided region, but the present invention is not limited to this. In the present invention, it is sufficient to determine whether or not there is variation in the degree of roughness of the surface S of the semiconductor wafer U, and for example, a histogram for each pixel value of the divided region may be generated, and based on the generated histogram, it may be determined whether or not there is variation in the degree of roughness in the surface S of the semiconductor wafer U shown in the appearance image 30. Also, an operator may look at the appearance image 30 shown on the display unit 6 and determine whether or not there is variation in the degree of roughness in the surface S of the semiconductor wafer U shown in the appearance image 30.

[0081] In the above first and second embodiments, examples have been shown in which the appearance images 10 to 30 are grayscale, but the present invention is not limited to this. In the present invention, the appearance images 10 to 30 may be color images. In this case, the threshold values ​​for each type of image may be determined using at least one value of the RGB values. [Explanation of symbols]

[0082] 1, 1a Processing section 2 Judgment section 3. Filter processing section 4. Detection section 5. Image acquisition section 6 Display section 10, 20, 30 Appearance images 10b, 20b Judgment images 10c First detection image (detection image) 10e Third detection image (detection image) 11, 21, 31, 41 Normal part 12, 13, 22, 32 Defective parts 20c Second detection image (detection image) 100, 100a Visual inspection device D. Defect S surface U Semiconductor wafer (object)

Claims

1. an image acquisition unit for acquiring an external image of an object; a processing unit that performs a detection filter process on the appearance image to generate a detection image for detecting a defect portion included in the appearance image, The processing unit is configured to perform the detection filter processing such that the detection sensitivity for the defective portion decreases as the degree of roughness of the surface of the object shown in the appearance image increases, thereby generating the detection image corresponding to the degree of roughness of the surface of the object shown in the appearance image.

2. 2. The appearance inspection apparatus according to claim 1, wherein the processing unit is configured to generate the detection image corresponding to the degree of roughness of the surface of the object shown in the appearance image by performing the detection filter processing that reduces a threshold value for detecting the defective portion contained in the appearance image so as to reduce the detection sensitivity for the defective portion as the degree of roughness of the surface of the object shown in the appearance image increases.

3. The appearance inspection device of claim 2, wherein the processing unit is configured to generate a judgment image for the appearance image by binarizing pixel values ​​based on a judgment threshold value which is a predetermined pixel value, and to judge the degree of roughness of the surface of the object shown in the appearance image based on the judgment image.

4. 4. The appearance inspection device according to claim 3, wherein the processing unit is configured to obtain a black spot ratio, which is the ratio of an area of ​​the judgment image occupied by areas of low binarized pixel values, and determine that a degree of roughness of the surface of the object depicted in the appearance image is small when the black spot ratio is smaller than a predetermined reference ratio, and determine that a degree of roughness of the surface of the object depicted in the appearance image is large when the black spot ratio is equal to or greater than the reference ratio.

5. The processing unit includes: generating a first detection image in which pixel values ​​of the appearance image are binarized based on a first pixel threshold value when it is determined that the degree of roughness of the surface of the object shown in the appearance image is small; The appearance inspection device of claim 4, configured to generate a second detection image for the appearance image by binarizing pixel values ​​based on a second pixel threshold value that is smaller than the first pixel threshold value when it is determined that the surface roughness of the object shown in the appearance image is large.

6. 6. The appearance inspection apparatus according to claim 5, wherein the processing unit is configured to further generate a third detection image in which pixel values ​​are binarized based on a third pixel threshold value that is greater than the first pixel threshold value.

7. The appearance image includes a plurality of regions, 2. The visual inspection apparatus of claim 1, wherein the processing unit is configured to generate a plurality of detection images corresponding to each of the plurality of regions in accordance with the degree of roughness of the surface of the object depicted in each of the plurality of regions, by performing a filter process that reduces the detection sensitivity for the defective portion contained in each of the plurality of regions as the degree of roughness of the surface of the object depicted in each of the plurality of regions increases.

8. The appearance inspection apparatus according to claim 1 , wherein the processing unit is configured to detect a position and a shape of the defective portion in the appearance image based on the detection image.

9. 7. The appearance inspection apparatus of claim 6, wherein the processing unit is configured to detect, as the defective portion in the appearance image, a portion of pixels in the first detection image that are smaller than the first pixel threshold value, a portion of pixels in the second detection image that are smaller than the second pixel threshold value, and a portion of pixels in the third detection image that are larger than the third pixel threshold value.

10. 9. The appearance inspection apparatus of claim 8, wherein the processing unit is configured to perform size filtering on the detection image so as to ignore the defect portion smaller than a predetermined size, thereby generating a size filtered image, and to detect the position and shape of the defect portion in the appearance image based on the size filtered image.

11. An image acquisition step of acquiring an appearance image of an object; and a detection image generation step of generating a detection image corresponding to the degree of surface roughness of the object shown in the appearance image by performing filter processing that reduces detection sensitivity for defects contained in the appearance image as the degree of surface roughness of the object shown in the appearance image increases.

Images