Surface process defect detection method and system for copper foil production

Abstract

The present application relates to the field of image enhancement technology, and specifically to a surface process defect detection method and system for copper foil production. The method comprises: collecting copper foil images during the production process; obtaining each image block in the copper foil image; performing edge detection on the copper foil image to obtain each edge line therein; determining the texture feature coefficient of each image block, constructing the illumination contrast coefficient of each image block, correcting the standard deviation parameter in the Gaussian filter, and using the Gaussian filter with the corrected parameters and the local Retinex algorithm to perform image enhancement on each image block; and performing surface defect detection on the copper foil using the image-enhanced copper foil image. This improves the accuracy of copper foil surface process defect detection.

Description

Technical Field

[0001] The present application relates to the field of image enhancement technology, and in particular to a surface process defect detection method and system for copper foil production. Background Art

[0002] With the rapid development of electronic information technology, copper foil, as an important basic material, is widely used in many fields such as printed circuit boards (PCBs) and lithium batteries. However, the quality requirements for copper foil are also very strict. Even minor defects can affect the performance and service life of the final product. Therefore, during the copper foil production process, it is crucial to ensure that the copper foil surface is defect-free. With the development of automation and image processing technology, automated surface process defect detection has gradually become one of the key technologies for improving product quality. This technology can monitor quality issues in the production process in real time and detect copper foil surface defects through advanced image processing technology.

[0003] However, in actual applications, the smoothness of the copper foil surface and its metallic properties make it extremely sensitive to changes in lighting conditions. Furthermore, lighting conditions at copper foil production sites can vary frequently, making them difficult to maintain consistent. These changes in lighting conditions can lead to noticeable differences in the color of the copper foil surface, which in turn affects image acquisition and defect identification. The Retinex algorithm, a commonly used image enhancement technology, can remove the effects of varying lighting conditions and improve the visual quality of images. However, when used for copper foil inspection, the Retinex algorithm is overly sensitive to lighting conditions due to the high reflectivity of the copper foil surface. Consequently, when the Retinex algorithm attempts to eliminate lighting effects, it is prone to light suppression, resulting in loss of image detail and difficulty in accurate defect detection. Summary of the Invention

[0004] In order to solve the above technical problems, the purpose of this application is to provide a surface process defect detection method and system for copper foil production. The technical solutions adopted are as follows:

[0005] In a first aspect, an embodiment of the present application provides a surface process defect detection method for copper foil production, the method comprising the following steps:

[0006] Collect images of copper foil during the production process;

[0007] Based on the grayscale values ​​of all pixels in the copper foil image, all pixels are divided into categories, and connected domains are extracted for pixels of the same category to obtain image blocks in the copper foil image;

[0008] Edge detection is performed on the copper foil image to obtain the edge lines therein; the texture feature coefficient of each image block is determined by the grayscale value change trend of the pixel points on the line connecting the center points of each image block and its adjacent image blocks, as well as the measured distance between each edge line in each image block and its image block center point;

[0009] Analyze the grayscale differences between all pixels in each image block and all pixels in the copper foil image, as well as the discreteness of the grayscale values ​​of the pixels in each image block. Combined with the differences in the texture feature coefficients of each image block and its adjacent image blocks, determine the illumination contrast coefficient of each image block.

[0010] Based on the proportion of the illumination contrast coefficient of each image block in all image blocks, the standard deviation parameter of the Gaussian filter is corrected. The Gaussian filter with the corrected parameters and the local Retinex algorithm are used to perform image enhancement on each image block. The surface defects of the copper foil are detected based on the enhanced copper foil image.

[0011] In one embodiment, determining the texture feature coefficient includes:

[0012] Calculate the mean of the metric distances between all edge lines in each image block and the center point of its image block, which is recorded as the first mean; connect each image block with the center point of its adjacent image block by a straight line, and perform linear fitting and curve fitting on the grayscale values ​​of all pixels on the straight line respectively;

[0013] The texture feature coefficient is determined based on the numerical distribution of the derivative of the function corresponding to the fitting curve and the slope of the fitting line in combination with the first mean.

[0014] In one embodiment, the further determination of the texture feature coefficient includes:

[0015] Substituting the grayscale values ​​of all pixels on the straight line into the derivative of the function corresponding to the fitting curve, obtaining the number of pixel grayscale values ​​whose derivative value is 0, and calculating the ratio of the absolute value of the slope of the fitting line to the number, which is recorded as a first ratio;

[0016] The texture feature coefficient is determined by combining the first ratio and the first mean, wherein the texture feature coefficient is positively correlated with the first ratio and the first mean.

[0017] In one embodiment, the mean of the first ratios corresponding to each image block and all its adjacent image blocks is calculated and recorded as the second mean, and the texture feature coefficient is the product of the first mean and the second mean.

[0018] In one embodiment, the metric distance between each edge line in each image block and the center point of the image block is: the average of the metric distances between all pixel points on each edge line in each image block and the center point of the image block.

[0019] In one embodiment, determining the illumination contrast coefficient includes:

[0020] Calculate the difference between the grayscale mean of all pixels in each image block and the grayscale mean of all pixels in the copper foil image, which is recorded as the first difference; and calculate the difference between the texture feature coefficients of each image block and its adjacent image blocks, which is recorded as the second difference;

[0021] determining a square root of a cumulative sum of the second differences between each image block and all its adjacent image blocks;

[0022] The illumination contrast coefficient is positively correlated with the first difference and the degree of dispersion, and negatively correlated with the square root result.

[0023] In one embodiment, a multiplication result of the first difference and the discreteness is calculated, and the illumination contrast coefficient is a ratio of the multiplication result to the square root result.

[0024] In one embodiment, the modifying of the standard deviation parameter in the Gaussian filter based on the proportion of the illumination contrast coefficient of each image block in all image blocks includes:

[0025] Calculating a ratio of the illumination contrast coefficient of each image block to the sum of the illumination contrast coefficients of all image blocks in the copper foil image, recording the ratio as a second ratio, using the second ratio as an exponent of an exponential function with a natural constant as a base, and using the calculation result of the exponential function as the illumination suppression weight of each image block;

[0026] The illumination suppression weight of each image block is used to correct the standard deviation parameter in the Gaussian filter.

[0027] In one embodiment, the product of a preset initial Gaussian filter standard deviation and the illumination suppression weight of each image block is used as a standard deviation parameter in the Gaussian filter when performing image enhancement on each image block using a local Retinex algorithm.

[0028] In a second aspect, an embodiment of the present application also provides a surface process defect detection system for copper foil production, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of any one of the above methods when executing the computer program.

[0029] This application has at least the following beneficial effects:

[0030] This application divides all pixels in the copper foil image into categories based on their grayscale values, extracts connected domains for pixels of the same type, and obtains image blocks in the copper foil image, thereby improving the accuracy of locating copper foil surface defects and enhancing the ability to perform refined analysis of copper foil surface defects. Furthermore, the texture feature coefficient of each image block in the copper foil image is determined, thereby improving the discrimination of complex textures on the copper foil surface, facilitating targeted identification of different defects on the copper foil surface, and improving the accuracy of defect detection. The illumination contrast coefficient of each image block is obtained, thereby solving the problem of error in defect detection caused by uneven illumination intensity, and significantly improving the discrimination between dark area defects and highlight area defects. The visibility of the copper foil surface defects is improved, and the interference of different lighting conditions on the copper foil surface defect detection is reduced; based on the proportion of the lighting contrast coefficient of each image block in all image blocks, the standard deviation parameter in the Gaussian filter is corrected, and the image enhancement of each image block is performed using the Gaussian filter with the corrected parameters and the local Retinex algorithm; the Retinex algorithm is made to focus more on the defect area, which optimizes the pertinence of image enhancement and is beneficial to retaining the defect details in the copper foil image, thereby enhancing the robustness of the local Retinex algorithm for image enhancement. The copper foil surface defect detection is performed on the copper foil image after image enhancement, thereby improving the accuracy and efficiency of defect detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0032] Figure 1 A flowchart of a method for detecting surface process defects in copper foil production according to an embodiment of the present application;

[0033] Figure 2 Flowchart for determining the standard deviation of a Gaussian filter. DETAILED DESCRIPTION

[0034] To further illustrate the technical means and effectiveness of this application to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effectiveness of the surface process defect detection method and system for copper foil production proposed in this application. In the following description, different references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0035] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0036] The specific scheme of the surface process defect detection method and system for copper foil production provided by this application is described in detail below with reference to the accompanying drawings.

[0037] See also Figure 1 , which shows a flowchart of a surface process defect detection method for copper foil production provided by an embodiment of the present application, the method comprising the following steps:

[0038] S1, collects copper foil images during the production process and performs preprocessing.

[0039] This embodiment uses a high-precision CCD industrial camera to capture image data from the copper foil production process, employing a high-brightness LED module as the light source. The captured copper foil image is then used as input, and a denoising algorithm is applied to remove image noise, outputting the denoised copper foil image. This embodiment employs a mean filter, a well-known technique. Implementers are free to choose other feasible denoising algorithms, such as median filtering or bilateral filtering, and this embodiment does not limit this approach.

[0040] S2, dividing all pixels in the copper foil image into categories based on their grayscale values, extracting connected domains for pixels of the same category, and obtaining image blocks in the copper foil image.

[0041] Copper foil is an important basic material, and its quality control is very strict. Even tiny defects on the copper foil have a significant impact on the performance and life of the product. Therefore, in the copper foil production of modern technology, industrial vision is introduced to ensure that the quality of the copper foil can be accurately detected during efficient production. Due to the characteristics of copper foil, the requirements for lighting conditions in the process of defect detection based on industrial vision are relatively high. Although controllable light sources are often introduced to ensure the similarity of lighting conditions as much as possible, in the actual production process, the errors caused by lighting conditions are still difficult to completely eliminate. Therefore, this embodiment uses Retinex image enhancement technology to further eliminate copper foil defect detection errors caused by lighting. However, when eliminating errors caused by differences in lighting conditions, it is inevitable to balance image details and light suppression intensity.

[0042] In the defect detection of copper foil surface, first of all, it is necessary to analyze the texture features in the copper foil image. Since different surface defects correspond to different texture features, the defects of the copper foil surface can be preliminarily measured to a certain extent based on the texture features.

[0043] The most common defects in copper foil are black and yellow spots. Black spots appear as circular black spots on the surface of the copper foil, shaped like the result of liquid diffusion and corrosion. The black spots exhibit a distinct step-like pattern from the inside out, with the color becoming darker as they approach the center. The center of the black spot is smooth and free of foreign matter. Yellow spots, on the other hand, show visible intrusion of foreign matter, making the center of the yellow spot appear highly uneven, with poor smoothness and a significant shift in color distribution.

[0044] In this embodiment, the denoised copper foil image is converted to a corresponding grayscale image using color space conversion. The grayscale value of each pixel in the grayscale image is used as input, and the natural breakpoint method is used to output the classification result of the grayscale value, that is, to divide all the pixels into categories. The optimal number of categories set by the natural breakpoint method is determined by the elbow method. Color space conversion, natural breakpoint method, and elbow method are all well-known technologies, and the specific process will not be described in detail. Then, the grayscale image is used as input, and the region growing algorithm is used to extract the connected domains in the grayscale image, wherein the similarity criterion is determined by the classification result of the grayscale value. If the grayscale value of the pixel point is in the same category, it is divided into the same connected domain. The connected domains divided in the grayscale image are recorded as image blocks. The region growing algorithm is an existing well-known technology, and the specific process will not be described in detail.

[0045] S3, performing edge detection on the copper foil image to obtain each edge line therein; determining the texture feature coefficient of each image block through the grayscale value change trend of the pixel points on the line connecting each image block and the center point of the adjacent image block, and the measured distance between each edge line in each image block and the center point of the image block.

[0046] The Canny edge detection algorithm is used to detect edges in the grayscale image and obtain edge lines. The Canny edge detection algorithm is a well-known technique, and implementers may choose other feasible existing edge detection algorithms; this embodiment does not limit this. For each image block segmented from the grayscale image, the minimum bounding rectangle of each image block is determined, and the intersection of the diagonals of the minimum bounding rectangle is used as the center point of the corresponding image block.

[0047] According to the characteristics of black spot defects and yellow spot defects on the copper foil surface, the texture features in the copper foil image are extracted and the texture feature coefficient is calculated to measure the texture change of the copper foil surface. Specifically:

[0048] Image blocks For example, assume that the image block With image blocks Adjacent, take image blocks With image blocks The grayscale value of each pixel on the straight line connecting the center points of the two image blocks, and along the image blocks Point to the image block The direction of the straight line forms the grayscale value sequence between the two image blocks. Then, using the elements in the grayscale value sequence as the vertical coordinates and the corresponding indexes of the elements as the horizontal coordinates, polynomial fitting and straight line fitting are performed, outputting the fitted curve function and the straight line equation. Polynomial fitting and straight line fitting are well-known techniques, and the specific processes are not repeated here.

[0049] The metric distance between each edge line in image block i and the center point of image block i is calculated. Specifically, the average of the metric distances between all pixels on each edge line in image block i and the center point of image block i is calculated. In this embodiment, the metric distance is calculated using Euclidean distance. The implementer may choose other existing feasible metric distance calculation methods.

[0050] Based on the above analysis, the texture feature coefficient of each image block is calculated to measure the texture change of the copper foil surface. The expression is:

[0051] Where, is an image block The texture feature coefficient of is an image block The number of edge lines in Is the image block The number of adjacent image blocks, is an image block Middle Edge lines and image blocks The metric distance between the center points of For image blocks The adjacent The slope of the fitting straight line of the gray value sequence between the image blocks is obtained, and the derivative function of the fitting curve function of the gray value sequence between the image block i and its adjacent t-th image block is obtained. is the number of grayscale value sequences between image block i and its adjacent t-th image block that take the value 0 in the derivative function. Recorded as the first mean, Recorded as the first ratio, It should be noted that the adjacent image blocks in this embodiment refer to image blocks with a common boundary between the two image blocks.

[0052] It should be understood that when the image block When the defect is black, the center of the defect is smooth and shows obvious step-type changes from the inside to the outside. The closer to the center of the defect, the smaller the gray value, so the image block The color change of the adjacent image blocks has a more obvious monotonic trend, and there will be no frequent color fluctuations. That is, the fewer the number of pixel gray values ​​whose gray value sequence takes the value of 0 in the derivative function, and the more the edge lines are distributed at a distance from the center of the image block, the larger the texture feature coefficient will be. On the contrary, when the image block When corresponding to yellow spot defects, the texture feature coefficient is smaller.

[0053] S4, analyzing the grayscale differences between all pixels in each image block and all pixels in the copper foil image, as well as the discreteness of the grayscale values ​​of the pixels in each image block, and determining the illumination contrast coefficient of each image block in combination with the differences in texture feature coefficients between each image block and its adjacent image blocks.

[0054] To ensure detail in the enhanced copper foil image and facilitate surface defect detection, the intensity of the light suppression must be adjusted. Traditional image enhancement methods apply uniform adjustments to the entire image. This approach results in significant loss of image detail in defective areas if the copper foil has defects, making it difficult to further improve defect detection accuracy. Therefore, this embodiment utilizes a more adaptable light suppression intensity adjustment method, eliminating errors caused by fluctuating lighting conditions while retaining sufficient detail for further defect detection.

[0055] For black spot defects on the copper foil surface, their shape is like the diffusion of liquid material, and the defect color distribution shows a step-like change, while the center of the defect appears to be a smooth surface. This means that the central area of ​​the black spot defect will still be affected by strong changes in lighting conditions. In the part where the defect shows a step-like change, it is necessary to pay attention to its specific details to avoid ignoring these step-like changes and causing the defect to be misjudged. Therefore, for black spot defects, strong light suppression can be used to eliminate the impact of different lighting conditions. For yellow spot defects on the copper foil surface, strong unevenness begins from the center area and the smoothness is poor. To ensure the recognition accuracy of yellow spot defects, weaker light suppression is required for the yellow spot defect area on the copper foil surface to ensure image details.

[0056] Finally, when performing illumination suppression on copper foil, it is also necessary to consider the contrast between different regions in the copper foil image and the global image, as well as the degree of variation in texture details in each region. If the local region shows a large contrast with the global image, or the texture detail features in a region are more complex and varied, it means that such regions contain more information. Therefore, weaker illumination suppression is also required to preserve the image detail features.

[0057] Based on the above analysis, this embodiment calculates the illumination contrast coefficient of each image block to measure the reflectivity difference of different areas on the copper foil surface, which in turn helps in subsequent illumination suppression adjustment. The specific expression is:

[0058] Where, is an image block The illumination contrast coefficient, 、 They are image blocks , the mean grayscale value of all pixels in the copper foil grayscale image, is an image block The discrete degree of grayscale values ​​of all pixels in Is the image block The number of adjacent image blocks, 、 They are image blocks , and image blocks The adjacent The texture feature coefficients of the image blocks. Recorded as the first difference, Recorded as the second difference.

[0059] It should be noted that the difference represents the degree of difference between two variables, which can be calculated specifically by using the absolute value of the difference, the square of the difference, the ratio, etc. The degree of dispersion can be calculated specifically by using the variance, standard deviation, coefficient of variation, etc. The implementer can choose it at his / her discretion. This embodiment does not impose any restrictions on this. This embodiment uses variance as the calculation method for the degree of dispersion.

[0060] It can be understood that for a normal image block, the grayscale value of the image block has a high similarity with the global grayscale value, and the grayscale value within the image block changes little. At the same time, the texture features within the image block are relatively regular, that is, is large, so the illumination contrast coefficient is relatively small at this time; for the image block corresponding to the black spot defect, the grayscale value in the image block is quite different from the global one, but the black spot defect is relatively smooth, that is, the grayscale value fluctuates little, so the corresponding illumination contrast coefficient increases; for the yellow spot defect, the grayscale value in the image block is quite different from the global one, and the grayscale value in the image block fluctuates greatly, so the illumination contrast coefficient is relatively the largest.

[0061] S5, based on the proportion of the illumination contrast coefficient of each image block in all image blocks, corrects the standard deviation parameter in the Gaussian filter, uses the Gaussian filter with the corrected parameters and the local Retinex algorithm to perform image enhancement on each image block; and uses the enhanced copper foil image to detect surface defects of the copper foil.

[0062] In image enhancement of copper foil surfaces, different lighting conditions have a significant impact on copper foil surface defect detection. Therefore, this embodiment eliminates the influence of lighting conditions through Retinex image enhancement technology. At the same time, when eliminating the influence of lighting conditions, in order to achieve a balance between lighting suppression and image details, this embodiment adopts a local Retinex algorithm to adaptively adjust the lighting suppression intensity of different regions. Based on the lighting contrast coefficient of each image block, an adaptive lighting suppression weight is determined, and the lighting suppression intensity of each image block is adaptively adjusted. This achieves the goal of minimizing the impact of lighting condition changes while retaining more image detail features, improving the quality of image enhancement, and enhancing the accuracy of surface defect detection in copper foil production.

[0063] Among them, the illumination suppression weight of each image block is determined, and the expression is:

[0064] Where, is an image block The illumination suppression weight, 、 They are image blocks , image blocks The illumination contrast coefficient, is the number of image blocks in the copper foil grayscale image, is an exponential function with a natural constant as its base. Recorded as the second ratio.

[0065] It can be understood that when the image block has fewer texture features and is more similar to the global image, its smoothness may be more affected by lighting conditions. At this time, a greater illumination suppression intensity is required to increase the strength of image enhancement, that is, the greater the illumination suppression weight; conversely, a smaller illumination suppression intensity is required to ensure the retention of image details, that is, the smaller the illumination suppression weight.

[0066] In this embodiment, the initial Gaussian filter standard deviation is set to ,Will The product of the illumination suppression weight of each image block is used as the Gaussian filter standard deviation when the local Retinex algorithm is used to enhance each image block, thereby enhancing each image block in the copper foil image and obtaining an enhanced copper foil image. The implementer can set it according to the actual situation, and this embodiment does not limit it. The local Retinex algorithm is a well-known technology, and the specific process is not described in detail. The Gaussian filter standard deviation determination flow chart is as follows: Figure 2 shown.

[0067] Accordingly, this embodiment collects the copper foil production process A copper foil image will be enhanced using the above method. A copper foil image is used as sample data to train a machine learning algorithm. The trained machine learning algorithm is then used to perform defect detection on the copper foil image, obtaining detection results for various defects and implementing defect detection for copper foil surface processing. In this embodiment, the machine learning algorithm uses the K-nearest neighbor algorithm, a well-known technique. The specific process is not detailed here. Implementers can choose other feasible defect detection algorithms, such as template matching or deep learning, and this embodiment does not impose any restrictions on these algorithms.

[0068] Based on the same inventive concept as the above-mentioned method, an embodiment of the present application also provides a surface process defect detection system for copper foil production, comprising a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of any one of the above-mentioned surface process defect detection methods for copper foil production are implemented.

[0069] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0070] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0071] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A surface process defect detection method for copper foil production, characterized in that: The method comprises the following steps: Collect images of copper foil during the production process; Based on the grayscale values ​​of all pixels in the copper foil image, all pixels are divided into categories, and connected domains are extracted for pixels of the same category to obtain image blocks in the copper foil image; Edge detection is performed on the copper foil image to obtain the edge lines therein; the texture feature coefficient of each image block is determined by the grayscale value change trend of the pixel points on the line connecting the center points of each image block and its adjacent image blocks, as well as the measured distance between each edge line in each image block and its image block center point; Analyze the grayscale differences between all pixels in each image block and all pixels in the copper foil image, as well as the discreteness of the grayscale values ​​of the pixels in each image block. Combined with the differences in the texture feature coefficients of each image block and its adjacent image blocks, determine the illumination contrast coefficient of each image block. Based on the proportion of the illumination contrast coefficient of each image block in all image blocks, the standard deviation parameter of the Gaussian filter is corrected. The Gaussian filter with the corrected parameters and the local Retinex algorithm are used to perform image enhancement on each image block. The surface defects of the copper foil are detected based on the enhanced copper foil image.

2. The surface process defect detection method for copper foil production according to claim 1, characterized in that: The determination of the texture feature coefficient includes: Calculate the mean of the metric distances between all edge lines in each image block and the center point of its image block, which is recorded as the first mean; connect each image block with the center point of its adjacent image block by a straight line, and perform linear fitting and curve fitting on the grayscale values ​​of all pixels on the straight line respectively; The texture feature coefficient is determined based on the numerical distribution of the derivative of the function corresponding to the fitting curve and the slope of the fitting line in combination with the first mean.

3. The surface process defect detection method for copper foil production according to claim 2, characterized in that: The further determination of the texture feature coefficient includes: Substituting the grayscale values ​​of all pixels on the straight line into the derivative of the function corresponding to the fitting curve, obtaining the number of pixel grayscale values ​​whose derivative value is 0, and calculating the ratio of the absolute value of the slope of the fitting line to the number, which is recorded as a first ratio; The texture feature coefficient is determined by combining the first ratio and the first mean, wherein the texture feature coefficient is positively correlated with the first ratio and the first mean.

4. The surface process defect detection method for copper foil production according to claim 3, characterized in that: The mean of the first ratios corresponding to each image block and all its adjacent image blocks is calculated and recorded as a second mean. The texture feature coefficient is the product of the first mean and the second mean.

5. The surface process defect detection method for copper foil production according to claim 1, characterized in that: The metric distance between each edge line in each image block and the center point of the image block is: the average of the metric distances between all pixel points on each edge line in each image block and the center point of the image block.

6. The surface process defect detection method for copper foil production according to claim 1, characterized in that: Determining the illumination contrast coefficient includes: Calculate the difference between the grayscale mean of all pixels in each image block and the grayscale mean of all pixels in the copper foil image, which is recorded as the first difference; and calculate the difference between the texture feature coefficients of each image block and its adjacent image blocks, which is recorded as the second difference; determining a square root of a cumulative sum of the second differences between each image block and all its adjacent image blocks; The illumination contrast coefficient is positively correlated with the first difference and the degree of dispersion, and negatively correlated with the square root result.

7. The surface process defect detection method for copper foil production according to claim 6, characterized in that: A multiplication result of the first difference and the discreteness is calculated, and the illumination contrast coefficient is a ratio of the multiplication result to the square root result.

8. The surface process defect detection method for copper foil production according to claim 1, characterized in that: The method of correcting the standard deviation parameter in the Gaussian filter based on the proportion of the illumination contrast coefficient of each image block in all image blocks includes: Calculating a ratio of the illumination contrast coefficient of each image block to the sum of the illumination contrast coefficients of all image blocks in the copper foil image, recording the ratio as a second ratio, using the second ratio as an exponent of an exponential function with a natural constant as a base, and using the calculation result of the exponential function as the illumination suppression weight of each image block; The illumination suppression weight of each image block is used to correct the standard deviation parameter in the Gaussian filter.

9. The surface process defect detection method for copper foil production according to claim 8, characterized in that: The product of the preset initial Gaussian filter standard deviation and the illumination suppression weight of each image block is used as the standard deviation parameter in the Gaussian filter when the local Retinex algorithm is used to perform image enhancement on each image block.

10. A surface process defect detection system for copper foil production, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 9 are implemented.

Images