A wafer surface smoothness detection method and system for visual inspection machine

Abstract

The present invention relates to the field of image data processing technology, and in particular to a method and system for detecting the smoothness of a wafer surface for a visual inspection machine. The method comprises: obtaining a grayscale image of a wafer surface image; dividing the grayscale image into a plurality of blocks; determining a first qualified coefficient of each block; obtaining edge pixel points within each block; determining a second qualified coefficient of each block; and clustering the grayscale image using a Naive hierarchical clustering algorithm to identify smooth areas in the grayscale image and detect the smoothness of the wafer surface. By introducing the first qualified coefficient and the second qualified coefficient, the present invention comprehensively considers multiple factors within the block, and can more comprehensively and meticulously characterize the characteristics of the block, thereby effectively distinguishing between rough areas and etched texture areas, reducing the interference of the etched texture area on the judgment of the rough area, and improving the accuracy of clustering, thereby more accurately identifying the rough area and ensuring the accuracy of the wafer surface smoothness detection.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing, and in particular to a method and system for detecting the surface smoothness of a wafer used in a visual inspection machine. Background Art

[0002] With the continuous advancement of emerging technologies such as artificial intelligence, 5G, the Internet of Things, and quantum computing, the semiconductor industry, as the foundation supporting these technologies, is experiencing unprecedented rapid development. Wafer smoothness inspection, the core of semiconductor products, is crucial for ensuring efficient manufacturing. Traditional manual visual inspection is inefficient and lacks accuracy, necessitating an automated, high-precision inspection technology.

[0003] An existing method for detecting wafer surface smoothness is the Naive Hierarchical Clustering Algorithm. The Naive Hierarchical Clustering Algorithm does not require a predefined number of clusters and can effectively handle complex and irregular data, making it an effective tool for detecting wafer surface smoothness.

[0004] However, when using the Naive hierarchical clustering algorithm for image processing, to reduce computational effort and improve algorithm efficiency, the image is divided into several equal-sized image blocks before clustering begins. Each image block is used as an initial cluster for subsequent clustering. During the clustering process, clusters are continuously merged based on the distance between them until only one cluster remains within the cluster set, resulting in a clustering result. The distance between two clusters is often quantified as a characteristic factor by analyzing the grayscale values ​​of each image block. However, when inspecting wafer surface smoothness, image blocks of potential rough areas can have similar grayscale values ​​to image blocks of some etched texture areas on the wafer surface. The image blocks of the etched texture areas can significantly affect the judgment of potential rough areas, resulting in a reduction in the accuracy of the Naive hierarchical clustering algorithm's clustering results, which in turn affects the accuracy of wafer surface smoothness inspection. Summary of the Invention

[0005] In order to solve the problem that the rough area on the wafer surface has a similar grayscale value expression as the etched texture area on the wafer surface, the etched texture area will greatly affect the judgment of the potential rough area, resulting in a decrease in the accuracy of clustering using the Naive hierarchical clustering algorithm, thereby affecting the accuracy of wafer surface smoothness detection, the present invention provides a wafer surface smoothness detection method and system for a visual inspection machine.

[0006] In a first aspect, the present invention provides a wafer surface smoothness detection method for a visual inspection machine, which adopts the following technical solution:

[0007] A wafer surface smoothness detection method for a visual inspection machine includes: obtaining a grayscale image of a wafer surface image; dividing the grayscale image into a plurality of equal blocks, determining a first qualification coefficient for each block based on the median of the grayscale values ​​of all pixels in the grayscale image, the mean of the grayscale values ​​of all pixels in the block, and the number of mode values ​​of the grayscale values ​​of all pixels in the block; recording pixels in the block with a gradient amplitude change in any neighborhood direction as edge pixels; in response to the block containing edge pixels, determining a second qualification coefficient for the block based on the first qualification coefficient, the number of edge pixels in the block, the number of pixels within a minimum circumscribed rectangle of all edge pixels in the block, and the information entropy of the gradient amplitudes corresponding to all edge pixels in the block; otherwise, assigning the first qualification coefficient of the block to the second qualification coefficient of the block; and clustering the grayscale image using a Naive hierarchical clustering algorithm based on the second qualification coefficient to obtain a dendrogram of the Naive hierarchical clustering algorithm to identify smooth areas in the grayscale image and implement wafer surface smoothness detection.

[0008] The beneficial effects are: by utilizing multiple features, such as the median, mean, and mode value number, to determine the first qualified coefficient of the block, it is possible to effectively distinguish between the surface roughness area and the etched texture area, and avoid misjudgment caused by the similarity of the grayscale values ​​between the two; by judging whether the block contains edge pixels and calculating the number of edge pixels, the minimum circumscribed rectangle and other features, it is possible to accurately identify areas with obvious unevenness or defects on the surface, and further improve the accuracy of smoothness detection; marking the pixels with gradient amplitude changes and calculating the information entropy value within the block can strengthen the processing of image texture complexity, reduce the influence of noise and non-target areas on the clustering algorithm, and thus improve the accuracy of detection.

[0009] Furthermore, the grayscale image is a grayscale image of the wafer surface image obtained by grayscale processing the collected wafer surface image using a weighted average method.

[0010] Furthermore, the grayscale image is divided into a plurality of blocks, including: taking any corner of the grayscale image as a reference point, dividing the grayscale image into a plurality of blocks. If there are remaining areas that cannot be evenly divided into In the block, the order of rows first and columns is followed and the size of the remaining area is divided, including: pre-dividing the remaining area according to the side lengths of adjacent blocks to obtain a number of pre-divided areas, and in response to the row length and column length of the pre-divided area being less than , merge the pre-divided area into the adjacent block, otherwise, treat the pre-divided area as a block; where, is the side length of the preset block, The minimum side length of the preset block.

[0011] Furthermore, the first qualification coefficient satisfies: Where, For the The first qualified coefficient of each block, is the median of the grayscale values ​​of all pixels in the grayscale image. For the The mean grayscale value of all pixels in a block, For the The number of mode values ​​among the gray values ​​of all pixels in a block, is a hyperparameter, is the absolute value symbol.

[0012] The beneficial effect is that by calculating the difference between the median of all pixels in the grayscale image and the mean of the grayscale values ​​of the pixels in each block, and then combining the number of mode values, the anomalies and differences in the grayscale distribution within the block can be captured more accurately, making it more sensitive when identifying non-smooth areas and reducing misjudgments caused by subtle changes in surface texture.

[0013] Furthermore, the gradient amplitude is calculated using a Sobel operator.

[0014] Furthermore, the gradient amplitude is calculated using a Canny operator.

[0015] Furthermore, the second qualification coefficient satisfies: Where, For the The second qualified coefficient of the block, For the The number of edge pixels in a block, For the The number of pixels within the minimum bounding rectangle of all edge pixels in a block, For the The first qualified coefficient of each block, For the The information entropy value of the gradient amplitude corresponding to all edge pixels in a block, is the standard normalization function.

[0016] The beneficial effects are: by calculating the second qualification coefficient of each block, the distribution of edge pixels within the image block can be described more accurately, while taking into account the number of pixels in the minimum circumscribed rectangle and the entropy value of the gradient amplitude, the details of the edge information in the image can be better captured; the second qualification coefficient can effectively enhance the robustness of the algorithm to image noise and uneven lighting by combining the first qualification coefficient and the entropy value of the gradient amplitude; by introducing a standard normalization function and normalizing the second qualification coefficient of the block, the deviation caused by differences in image size or differences in the blocking method can be eliminated, thereby making the quality assessment between multiple blocks more consistent.

[0017] Furthermore, clustering the grayscale image using the Naive hierarchical clustering algorithm according to the second qualified coefficient to obtain a dendrogram of the Naive hierarchical clustering algorithm includes: using the difference in the second qualified coefficients between the blocks as the distance between the blocks when the Naive hierarchical clustering algorithm processes the image; clustering the grayscale image using the Naive hierarchical clustering algorithm in a single-link manner to generate a clustering result, i.e., a dendrogram of the Naive hierarchical clustering algorithm.

[0018] The beneficial effects are: by using the second qualification coefficient to measure the differences between image blocks, the similarity between different areas of the image can be described more accurately, so that the Naive hierarchical clustering algorithm can perform more refined clustering based on real image features; compared with traditional clustering methods, the Naive hierarchical clustering algorithm clusters through a single link method, so that the image can obtain a hierarchical structure, i.e., a dendrogram, more quickly during processing, thereby improving the computational efficiency of the clustering process and reducing a large amount of computation.

[0019] Furthermore, the method of detecting the smoothness of the chip surface by identifying the smooth area in the grayscale image includes: cropping the dendrogram according to a preset cluster distance to obtain a plurality of cropped clusters; in response to the number of blocks in the cropped cluster being less than a preset roughness threshold, determining that the grayscale image area corresponding to the cropped cluster is a non-smooth area; in response to the ratio of the area of ​​all non-smooth areas to the area of ​​the grayscale image being greater than a preset abnormality threshold, determining that the chip surface quality is unqualified, and completing the detection of the chip surface smoothness.

[0020] The beneficial effects are: by cropping the grayscale image according to the dendrogram and judging the smoothness of the area according to the size of the cluster, the smooth and rough areas on the chip surface can be accurately distinguished. By comparing the number of blocks in the rough area with the preset roughness threshold, the surface irregularities or unevenness can be effectively identified, thereby improving the accuracy of detection; by calculating the ratio of the area of ​​the rough area to the area of ​​the entire image, the overall smoothness of the chip surface can be quickly judged.

[0021] In a second aspect, the present invention provides a wafer surface smoothness detection system for a visual inspection machine, which adopts the following technical solution:

[0022] A wafer surface smoothness detection system for a visual inspection machine includes: a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, the above-mentioned wafer surface smoothness detection method for a visual inspection machine is implemented.

[0023] By adopting the above technical solution, the above-mentioned chip surface smoothness detection method for a visual inspection machine is generated into a computer program and stored in a memory so as to be loaded and executed by a processor, thereby making a terminal device based on the memory and the processor for easy use.

[0024] The present invention has the following technical effects:

[0025] By introducing the first qualification coefficient and the second qualification coefficient, multiple factors such as the median, mean, mode number of pixel grayscale values ​​in the block, the number of edge pixels, the number of pixels in the minimum circumscribed rectangle of edge pixels, and the information entropy value of the gradient amplitude of edge pixels are comprehensively considered. This can more comprehensively and meticulously characterize the characteristics of the block, thereby effectively distinguishing the non-smooth area from the etched texture area, reducing the interference of the etched texture area on the judgment of the non-smooth area, and improving the accuracy of clustering. When clustering using the Naive hierarchical clustering algorithm, the second qualification coefficient is used as the distance measurement method between blocks. Since the second qualification coefficient integrates multiple information, it can be more accurate than clustering based on grayscale values ​​alone. Accurately reflecting the similarities and differences between blocks makes the clustering results more reasonable, and can more clearly distinguish smooth areas from non-smooth areas, avoiding erroneous clustering due to similar grayscale values, thereby improving the accuracy of clustering; the accuracy of clustering directly affects the accuracy of wafer surface smoothness detection. By accurately identifying non-smooth areas in grayscale images, the smoothness of the wafer surface can be more accurately judged. For visual inspection machines in industrial production, this means that wafers with surfaces that do not meet the requirements can be more effectively screened out, improving the reliability of product quality detection, and reducing the outflow of defective products or misjudgments due to inaccurate detection, thereby reducing production costs and improving production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is a flow chart of a method for detecting wafer surface smoothness in a visual inspection machine according to an embodiment of the present invention. DETAILED DESCRIPTION

[0027] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work shall fall within the scope of protection of the present invention.

[0028] The embodiment of the present invention discloses a method for detecting the surface smoothness of a wafer used in a visual inspection machine, referring to Figure 1 , including steps S1 to S6:

[0029] S1: Acquire a grayscale image of the wafer surface.

[0030] Specifically, the grayscale image is a grayscale image of the wafer surface image obtained by grayscale processing the wafer surface image collected by a high-definition camera using a weighted average method.

[0031] S2: Divide the grayscale image into several equal blocks.

[0032] Specifically, dividing the grayscale image into a number of equal blocks includes:

[0033] Taking any corner of the grayscale image as the reference point, the grayscale image is divided into several If there are remaining areas that cannot be evenly divided into In the blocks, the order of rows first and columns second is followed, and the remaining area is divided according to its size, including:

[0034] The remaining area is pre-divided according to the side lengths of adjacent blocks to obtain several pre-divided areas. In response to the row length and column length of the pre-divided area being less than , merge the pre-divided areas into adjacent blocks, otherwise, treat the pre-divided areas as one block;

[0035] in, is the side length of the preset block, The minimum side length of the preset block.

[0036] Implementers can set the side length and minimum side length based on specific implementation circumstances. For example, the side length is 15 and the minimum side length is 10.

[0037] S3: Determine the first qualified coefficient of the block.

[0038] It should be noted that by analyzing the grayscale value performance in each image block, the first qualified coefficient of each image block is obtained. Since most areas of the chip surface will have very close grayscale values, that is, the background area, and the potential non-smooth area will have grayscale changes due to the difference in light reflection, the grayscale value of the background area on the chip surface will be closer to the median grayscale value of the entire chip surface image relative to the potential non-smooth area; however, there will be some etched textures in the background area on the chip surface, and these etched texture areas and the background area belong to the normal area of ​​the chip surface. The etched texture areas and the background areas on the chip surface will have a high degree of grayscale value consistency; therefore, first analyze the difference between the grayscale value in an image block and the median grayscale value in the grayscale image of the chip surface image. The smaller the difference, the larger the first qualified coefficient. Then continue to analyze the consistency of the grayscale value in an image block. The greater the consistency, the larger the first qualified coefficient.

[0039] The first qualified coefficient of the block is determined according to the median of the grayscale values ​​of all pixels in the grayscale image, the mean of the grayscale values ​​of all pixels in the block, and the number of mode values ​​of the grayscale values ​​of all pixels in the block.

[0040] Specifically, the first qualification coefficient satisfies:

[0041] ;

[0042] Where, For the The first qualified coefficient of each block, is the median of the grayscale values ​​of all pixels in the grayscale image. For the The mean grayscale value of all pixels in a block, For the The number of mode values ​​among the gray values ​​of all pixels in a block, is a hyperparameter, is the absolute value symbol.

[0043] Implementers can set hyperparameters according to specific implementation conditions, for example, 0.01. The existence of hyperparameters is to prevent , the formula result becomes meaningless.

[0044] in, The smaller the The closer the gray value of a block is to the median of the gray value of the entire chip surface, the The larger the first qualified coefficient of each block, The larger the The greater the consistency of the gray value within a block, the greater the The closer the gray value of a block is to the median gray value of the entire chip surface, the greater the credibility is. The larger the first qualified coefficient of each block is, the greater the first qualified coefficient of each block is.

[0045] S4: Get edge pixels within the block.

[0046] Pixels with gradient amplitude changes in any neighborhood direction within a block are recorded as edge pixels, wherein the neighborhood directions are four neighborhood directions in the present invention.

[0047] S5: Determine the second qualification coefficient of the block.

[0048] It should be noted that the first qualified coefficient of the combined area of ​​the etched texture area and the background area on the chip surface will be close to the potential rough area, so it is also necessary to analyze the texture feature performance in each image block and combine the first qualified coefficient to obtain the second qualified coefficient of each image. The second qualified coefficient can better distinguish the potential rough area from the normal area (background area, etched texture area). The difference in texture feature expression between the potential rough area and the etched texture area is that the area composed of edge pixels in the etched texture area is relatively regular, while the area composed of edge pixels in the potential rough area is relatively irregular (this is due to the different causes of the two areas. The etched texture is produced according to consistent specifications, while the generated gray of the potential rough area is more random). However, there may also be some potential rough areas with relatively regular edge shapes. Then continue to analyze the consistency of the gradient change intensity of the pixels with gradient amplitude changes in an image block. If the consistency is poor, it is a potential rough area. Otherwise, it is an etched texture area. After the above feature analysis, if the texture feature expression in an image block is more consistent with the texture feature expression of the potential rough area, the second qualified coefficient corresponding to the image block will be smaller; if there is no edge pixel in an image block, then the image block does not have the above features, and the value of the first qualified coefficient of the image block can be assigned to the second qualified coefficient for use in subsequent steps.

[0049] In response to the presence of edge pixels in the block, the second qualified coefficient of the block is determined based on the first qualified coefficient, the number of edge pixels in the block, the number of pixels in the minimum circumscribed rectangle of all edge pixels in the block, and the information entropy value of the gradient amplitude corresponding to all edge pixels in the block; otherwise, the first qualified coefficient of the block is assigned to the second qualified coefficient of the block.

[0050] Specifically, the gradient amplitude is calculated using a Sobel operator.

[0051] In another embodiment, the gradient amplitude is calculated using a Canny operator.

[0052] Specifically, the second qualification coefficient satisfies:

[0053] ;

[0054] Where, For the The second qualified coefficient of the block, For the The number of edge pixels in a block, For the The number of pixels within the minimum bounding rectangle of all edge pixels in a block, For the The first qualified coefficient of each block, For the The information entropy value of the gradient amplitude corresponding to all edge pixels in a block, is the standard normalization function.

[0055] in, The larger the The more likely a block is to belong to the normal area of ​​the chip surface, the higher the gray value of the block. The larger the second qualified coefficient of each block, Indicates the The ratio of all pixels (edge ​​pixels) with gradient amplitude changes in the four neighborhood directions in a block to the area of ​​its minimum circumscribed rectangle. The smaller the value, the higher the gradient amplitude. The more irregular the shape of the region composed of edge pixels in a block, the more its texture characteristics conform to the texture characteristics of the potential non-smooth region. The second qualified coefficient of each block will be smaller; The larger the The worse the consistency of the gradient change intensity of the edge pixels in each block, the The more irregular the shape of the area composed of edge pixels in the block, the greater the credibility, which can explain the The more the texture features in a block are consistent with the texture features of the potential non-smooth area, the The smaller the second qualified coefficient of each block, the smaller the second qualified coefficient of each block.

[0056] S6: Use the Naive hierarchical clustering algorithm to cluster the grayscale image to identify the smooth area in the grayscale image and realize the smoothness detection of the wafer surface.

[0057] According to the second qualified coefficient, the grayscale image is clustered using the Naive hierarchical clustering algorithm to obtain a dendrogram of the Naive hierarchical clustering algorithm to identify smooth areas in the grayscale image and realize smoothness detection of the wafer surface.

[0058] Specifically, clustering the grayscale image using the Naive hierarchical clustering algorithm according to the second qualification coefficient to obtain a dendrogram of the Naive hierarchical clustering algorithm includes:

[0059] The difference in the second qualified coefficients between the blocks is used as the distance between the blocks when the Naive hierarchical clustering algorithm processes the image;

[0060] The grayscale image is clustered using the Naive hierarchical clustering algorithm in a single-link manner to generate a clustering result, namely a dendrogram of the Naive hierarchical clustering algorithm.

[0061] Specifically, the method of detecting the smoothness of the wafer surface by identifying the smooth area in the grayscale image includes:

[0062] The dendrogram is pruned according to a preset cluster distance to obtain a plurality of pruned clusters;

[0063] In response to the number of blocks in the cropped cluster being less than a preset roughness threshold, determining the grayscale image region corresponding to the cropped cluster as a non-smooth region;

[0064] In response to the ratio of the area of ​​all the non-smooth regions to the area of ​​the grayscale image being greater than a preset abnormal threshold, the wafer surface quality is determined to be unqualified, and the detection of the wafer surface smoothness is completed.

[0065] Implementers can set the cluster distance, rough threshold and abnormal threshold according to specific implementation conditions. For example, the cluster distance is 0.3, the rough threshold is 6, and the abnormal threshold is 0.03.

[0066] An embodiment of the present invention also discloses a wafer surface smoothness detection system for a visual inspection machine, comprising a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, a wafer surface smoothness detection method for a visual inspection machine according to the present invention is implemented.

[0067] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and will not be described in detail here.

[0068] The above are all preferred embodiments of the present invention, and are not intended to limit the scope of protection of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for detecting the surface smoothness of a wafer for a visual inspection machine, characterized in that: include: Acquire a grayscale image of the wafer surface image; Divide the grayscale image into several equal blocks; The first qualified coefficient of the block is determined based on the median of the grayscale values ​​of all pixels in the grayscale image, the mean of the grayscale values ​​of all pixels in the block, and the number of mode values ​​of the grayscale values ​​of all pixels in the block. The first qualified coefficient satisfies: Where, For the The first qualified coefficient of each block, is the median of the grayscale values ​​of all pixels in the grayscale image. For the The mean grayscale value of all pixels in a block, For the The number of mode values ​​among the gray values ​​of all pixels in a block, is a hyperparameter, is the absolute value symbol; The pixels with gradient amplitude changes in any neighborhood direction within the block are recorded as edge pixels; In response to the block containing edge pixels, determining a second qualification coefficient for the block based on the first qualification coefficient, the number of edge pixels in the block, the number of pixels in the minimum circumscribed rectangle of all edge pixels in the block, and the information entropy value of the gradient amplitude corresponding to all edge pixels in the block; otherwise, assigning the first qualification coefficient of the block to the second qualification coefficient of the block; The second qualification coefficient meets the following requirements: Where, For the The second qualified coefficient of the block, For the The number of edge pixels in a block, For the The number of pixels within the minimum bounding rectangle of all edge pixels in a block, For the The information entropy value of the gradient amplitude corresponding to all edge pixels in a block, is the standard normalization function; The difference in the second qualified coefficients between the blocks is used as the distance between the blocks for clustering when processing the image using the Naive hierarchical clustering algorithm, and a dendrogram of the Naive hierarchical clustering algorithm is obtained to identify smooth areas in the grayscale image and realize the smoothness detection of the wafer surface.

2. The wafer surface smoothness detection method for a visual inspection machine according to claim 1, characterized in that: The grayscale image is a grayscale image of the wafer surface image obtained by grayscale processing the collected wafer surface image using a weighted average method.

3. The wafer surface smoothness detection method for a visual inspection machine according to claim 1, characterized in that: The grayscale image is divided into a number of equal blocks, including: Taking any corner of the grayscale image as the reference point, the grayscale image is divided into several If there are remaining areas that cannot be evenly divided into In the blocks, the order of rows first and columns second is followed, and the remaining area is divided according to its size, including: The remaining area is pre-divided according to the side lengths of adjacent blocks to obtain several pre-divided areas. In response to the row length and column length of the pre-divided area being less than , merge the pre-divided areas into adjacent blocks, otherwise, treat the pre-divided areas as one block; in, is the side length of the preset block, The minimum side length of the preset block.

4. The wafer surface smoothness detection method for a visual inspection machine according to claim 1, characterized in that: The gradient amplitude is calculated using the Sobel operator.

5. The wafer surface smoothness detection method for a visual inspection machine according to claim 1, characterized in that: The gradient amplitude is calculated using the Canny operator.

6. The wafer surface smoothness detection method for a visual inspection machine according to claim 1, characterized in that: Cluster the blocks and obtain the dendrogram of the Naive hierarchical clustering algorithm, including: The grayscale image is clustered using the Naive hierarchical clustering algorithm in a single-link manner to generate a clustering result, namely a dendrogram of the Naive hierarchical clustering algorithm.

7. The wafer surface smoothness detection method for a visual inspection machine according to claim 1, characterized in that: The method of detecting the smoothness of the wafer surface by identifying the smooth area in the grayscale image includes: The dendrogram is pruned according to a preset cluster distance to obtain a plurality of pruned clusters; In response to the number of blocks in the cropped cluster being less than a preset roughness threshold, determining the grayscale image region corresponding to the cropped cluster as a non-smooth region; In response to the ratio of the area of ​​all the non-smooth regions to the area of ​​the grayscale image being greater than a preset abnormal threshold, the wafer surface quality is determined to be unqualified, and the detection of the wafer surface smoothness is completed.

8. A wafer surface smoothness detection system for a visual inspection machine, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a wafer surface smoothness detection method for a visual inspection machine according to any one of claims 1 to 7 is implemented.

Images