Copper foil width detection device and method based on visual identification algorithm

Abstract

The invention discloses a copper foil width detection device and method based on a visual identification algorithm, and the device comprises two cameras which are located at the two sides of a to-be-detected copper foil respectively, are fixed on a bridge, and are used for collecting the image data of the edges of the two sides of the copper foil; the driving device is used for conveying the copper foil and ensuring that two side edges of the copper foil stably move in the visual field of the camera; and the rear-end analysis server receives the image data from the camera, preprocesses the image data, calculates the width of the copper foil through a visual identification algorithm, and transmits a result to a display screen for displaying. According to the device for detecting the width of the copper foil through the visual identification algorithm, the width of the whole copper foil can be completely covered, the images of the edges of the two sides of the copper foil are shot through the camera, the width is measured through the image processing algorithm, and the device has the advantages of being non-contact, real-time and efficient.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial automated detection of copper foil width, and in particular to a copper foil width detection device and method based on a visual recognition algorithm. Background Art

[0002] In the copper foil production process, the width of the copper foil is an important quality indicator. In order to ensure that the width of each roll of copper foil meets the design requirements, the copper foil needs to be detected online. However, there are certain challenges in measuring the width of the copper foil. Traditional width detection methods usually rely on manual measurement or contact measurement equipment, which have some problems: first, manual measurement is prone to human errors and is inefficient; second, although contact measurement equipment can provide higher accuracy, its detection process is complicated and cannot achieve online real-time detection. In addition, contact measurement may also cause damage to the surface of the copper foil, affecting product quality.

[0003] With the development of industrial automation technology, non-contact detection methods based on visual recognition algorithms have gradually been applied to various industrial detection fields. This method uses a camera to capture copper foil images and uses image processing algorithms to measure the width. It has the advantages of non-contact, real-time and high efficiency. However, existing visual inspection systems usually use single-point measurement equipment, which can only detect a specific point of the copper foil and cannot fully cover the width of the entire copper foil, resulting in limitations in the measurement results and failure to reflect the width changes of the entire copper foil. At the same time, the single-point measurement system is easily affected by noise and light changes in the production environment, resulting in unstable measurement results and insufficient accuracy. Especially when the surface of the copper foil is smooth and reflective, the measurement error of the traditional visual system is large.

[0004] Patent document with application number 202410658166.6 discloses a copper foil size detection device, comprising: a movable seat, movably disposed on a support; an image acquisition component, disposed on the movable seat, the moving path of the image acquisition component passing through the copper foil, the image acquisition component being used to respectively acquire images of two edges of the copper foil's size to be measured; a position detection component, used to detect the position information of the image acquisition component; a processing component connected to the image acquisition component and the position detection component, the processing component calculating the distance value between the edge and the center of the image based on the image, calculating the displacement value of the image acquisition component between the acquisition positions of the two images based on the position information, and calculating the size to be measured based on the distance value and the displacement value.

[0005] The application scheme can calculate the measured size of the copper foil based on the collected image and the displacement value of the mobile base, without manual reading, which solves the problem of manual detection reading, which is time-consuming and labor-intensive, and has reading errors. However, it also has the following problems: it uses a mobile base, which makes it difficult to continuously collect accurate images of the edges of both sides of the copper foil, and it does not pre-process the image, making it difficult to calculate the accurate width value. Summary of the invention

[0006] In view of the above-mentioned problems, the purpose of the present invention is to provide a copper foil width detection device and method based on a visual recognition algorithm, which utilizes an image processing algorithm to perform non-contact, real-time, and efficient continuous width measurement of the entire copper foil.

[0007] The purpose of the present invention can be achieved by the following technical solution: A copper foil width detection device based on a visual recognition algorithm, comprising:

[0008] Two cameras are located on both sides of the copper foil to be inspected and fixed on the bridge to collect image data of the edges of both sides of the copper foil;

[0009] A drive device is used to transport the copper foil and ensure that the edges of the copper foil move smoothly within the camera's field of view;

[0010] The back-end analysis server receives image data from the camera, performs image data preprocessing, calculates the width of the copper foil through a visual recognition algorithm, and transmits the result to a display screen for display.

[0011] As a further solution of the present invention, it also includes:

[0012] The alarm device sends out an audible and visual alarm when the width of the copper foil calculated by the back-end analysis server exceeds the preset range.

[0013] A copper foil width detection method based on a visual recognition algorithm applied to the above device, the method comprising the steps of:

[0014] S1. Keep the edges of the copper foil moving smoothly within the field of view of the two cameras;

[0015] S2, using two cameras to collect image data of the edges of the copper foil on both sides, and sending them to the back-end analysis server;

[0016] S3, using the backend analysis server to perform image data preprocessing on the collected image data;

[0017] S4, calculating the width of the copper foil through a visual recognition algorithm, and transmitting the result to a display screen for display;

[0018] S5. When the calculated copper foil width exceeds the preset range, an audible and visual alarm is issued.

[0019] As a further solution of the present invention, the preprocessing of the image data in S3 includes: gray-scaling, denoising and image enhancement preprocessing of the image data.

[0020] As a further solution of the present invention, the image data is preprocessed into grayscale, including converting the captured color image into a grayscale image, and the formula is:

[0021] Gray=0.299×R+0.587×G+0.114×B (1)

[0022] Where R, G, and B are the red, green, and blue channels of the image, respectively.

[0023] As a further solution of the present invention, the image data is pre-processed for denoising, including using a Gaussian filter to smooth the image and reduce noise interference. The calculation formula is:

[0024]

[0025] Here, σ is the standard deviation of the Gaussian function.

[0026] As a further solution of the present invention, the image data is subjected to image enhancement preprocessing, including using adaptive histogram equalization to improve the contrast of the image and enhance the local detail features of the image, the steps are:

[0027] S31, dividing the image into several small areas, each of which has a size of M×N, and independently performing histogram equalization on each small area. The conversion function formula of the histogram equalization is:

[0028]

[0029] Where I is the pixel value of the small area, H(j) is the histogram, L is the number of gray levels, and M×N is the number of pixels in the small area;

[0030] S32, then for any bin in the histogram, if its frequency exceeds the threshold, the excess part is evenly distributed to all other bins;

[0031] S33, performing bilinear interpolation and merging on the equalization results of adjacent small blocks to obtain a final enhanced image.

[0032] As a further solution of the present invention, the visual recognition algorithm uses the Sobel operator to identify the edges of objects in the image, performs image edge detection, identifies the edge position of the copper foil, and uses morphological operations to improve the continuity of the edge.

[0033] As a further solution of the present invention, the method of using the Sobel operator to identify the edge of the copper foil in the image includes the steps of:

[0034] Use two convolution kernels to calculate the gradient information of the image in the horizontal and vertical directions respectively;

[0035] Among them, the convolution kernel gradient information G x Used to detect the horizontal edge of the image, G x Emphasize the grayscale change in the horizontal direction. x By performing convolution operation with the image, the gradient information of the image in the horizontal direction can be obtained;

[0036] Convolution kernel gradient information G y Used to detect the vertical edge of the image, G y Emphasize the grayscale change in the vertical direction. y By performing convolution operation with the image, the gradient information of the image in the vertical direction can be obtained;

[0037] According to the gradient information in the horizontal and vertical directions of the image, the image gradient size and gradient direction are obtained to detect the edge, where:

[0038] The image gradient size calculation formula is:

[0039]

[0040] The image gradient direction calculation formula is:

[0041]

[0042] Among them, G x is the gradient information in the horizontal direction, G y It is the vertical gradient information.

[0043] As a further solution of the present invention, the calculation of the width of the copper foil includes obtaining the actual distance by calculating the pixel distance between the edges of the copper foil on both sides of the image based on the edge detection result, wherein:

[0044] The pixel distance calculation formula is:

[0045] Width(in pixels)=|x2-x1| (6)

[0046] Among them, x1 and x2 are the boundary line detection results on both sides of the copper foil;

[0047] The conversion formula between actual distance and pixel distance is:

[0048] W = Width (in pixels) × d (7)

[0049] Where W is the actual distance and d is the conversion factor from pixel distance to actual distance.

[0050] Beneficial effects of the present invention:

[0051] 1. The present invention uses a visual recognition algorithm in a copper foil width detection device to fully cover the entire width of the copper foil. The camera continuously collects edge images of both sides of the copper foil and uses an image processing algorithm to measure the width. It has the advantages of being continuous, non-contact, real-time and efficient.

[0052] 2. The present invention performs preprocessing on the collected image, including denoising, graying and image enhancement. The preprocessed image can reduce noise interference and improve the accuracy of edge detection.

[0053] 3. The present invention uses a visual recognition algorithm and a Sobel operator to identify the edges of objects in an image, and uses the gradients of the image in the horizontal and vertical directions to perform edge detection. The detection position is accurate and is not easily affected by noise and light changes in the production environment. The measurement results are stable and highly accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a structural schematic diagram of a copper foil width detection device based on a visual recognition algorithm of the present invention;

[0055] Figure 2 It is a schematic diagram of the process of the copper foil width detection method based on the visual recognition algorithm of the present invention;

[0056] Figure 3 It is a schematic diagram of the principle of the copper foil width detection method based on the visual recognition algorithm of the present invention.

[0057] 110. Camera; 120. Bridge; 130. Driving device; 140. Display screen; 150. Back-end analysis server; 160. Copper foil; 170. Foil production machine; 180. Alarm device. DETAILED DESCRIPTION

[0058] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar symbols throughout represent the same or similar elements or elements with the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0059] like Figure 1 As shown, the present invention discloses a copper foil 160 width detection device based on a visual recognition algorithm, comprising:

[0060] Two cameras 110 are respectively located on two sides of the copper foil 160 to be inspected and fixed on the bridge 120 to collect image data of the edges of the copper foil 160 on both sides.

[0061] The driving device 130 is used to transport the copper foil 160 to ensure that the edges of both sides of the copper foil 160 move smoothly within the field of view of the camera 110 .

[0062] The back-end analysis server 150 receives the image data from the camera 110 , performs image data preprocessing, calculates the width of the copper foil 160 through a visual recognition algorithm, and transmits the result to the display screen 140 for display.

[0063] The alarm device 180 sends out an audible and visual alarm when the width of the copper foil 160 calculated by the back-end analysis server 150 exceeds a preset range.

[0064] The two cameras 110 are respectively located on both sides of the copper foil 160 to be inspected. The camera 110 can adopt a high-resolution industrial camera to ensure the accuracy and clarity of the image data; the camera 110 collects image data of the edges of the copper foil 160 on both sides. Each camera 110 is fixed on a bridge 120. The bridge 120 can be made of metal material with good rigidity and stability, ensuring that the camera 110 can stably capture the image data of the position of the edge of the copper foil 160; the bridge 120 is fixed above the foil production machine 170, ensuring that the camera 110 can stably capture the image data of the distance between the copper foil 160 and the edge.

[0065] The driving device 130 is used to transport the copper foil 160. The driving device 130 can use linear motors, belts, rollers and other devices to form a roller mechanism to ensure that the edge of the copper foil 160 moves smoothly within the field of view of the camera 110 and reduce deviation or jitter during movement.

[0066] The display screen 140 is used to display the images and detection results collected by the camera 110 in real time, which is convenient for operators to monitor and adjust; the back-end analysis server 150 has built-in image processing software, which pre-processes the image data collected by the camera 110, calculates the width of the copper foil 160 using a visual recognition algorithm, and transmits the result to the display screen 140; when the result exceeds the set value, the display screen 140 and the alarm device 180 can perform alarm operations at the same time.

[0067] like Figure 2 As shown, the present invention also discloses a method for detecting the width of the copper foil 160 based on a visual recognition algorithm using the above device, comprising the steps of:

[0068] S1, keep the edges of both sides of the copper foil 160 moving smoothly within the field of view of the two cameras 110;

[0069] S2, using two cameras 110 to collect image data of the edges of the copper foil 160 on both sides, and sending the image data to the back-end analysis server 150;

[0070] S3, using the backend analysis server 150 to perform image data preprocessing on the collected image data;

[0071] S4, calculating the width of the copper foil 160 by a visual recognition algorithm, and transmitting the result to the display screen 140 for display;

[0072] S5. When the calculated width of the copper foil 160 exceeds a preset range, an audible and visual alarm is issued.

[0073] like Figure 3 As shown, the preprocessing of the image data includes: graying, denoising and image enhancement preprocessing of the image data. The preprocessed image can reduce noise interference and improve the accuracy of edge detection.

[0074] The image data is preprocessed into grayscale, including converting the captured color image into a grayscale image. The formula is:

[0075] Gray=0.299×R+0.587×G+0.114×B (1)

[0076] Where R, G, and B are the red, green, and blue channels of the image, respectively.

[0077] The image data is preprocessed to remove noise, including using a Gaussian filter to smooth the image and reduce noise interference. The calculation formula is:

[0078]

[0079] Here, σ is the standard deviation of the Gaussian function.

[0080] Furthermore, the image data is subjected to image enhancement preprocessing, including using adaptive histogram equalization (CLAHE) to improve the contrast of the image and enhance the local detail features of the image. The steps are as follows:

[0081] First, the image is divided into several small areas, each of which is M×N in size. Histogram equalization is performed independently on each small area. The conversion function formula of histogram equalization is:

[0082]

[0083] Among them, I is the pixel value of the small block area, H(j) is the histogram, L is the number of gray levels, and M×N is the number of pixels in the small block area.

[0084] Then, for any bin in the histogram, if its frequency exceeds the threshold, the excess part will be evenly distributed to all other bins; this can prevent excessive amplification of certain gray levels and suppress noise.

[0085] Since the image is composed of multiple small blocks, in order to avoid the blocking effect, the equalization results of adjacent small blocks are then combined by bilinear interpolation to obtain the final enhanced image.

[0086] The preprocessed image data is subjected to a visual recognition algorithm, which includes using a Sobel operator to identify the edges of objects in the image, performing image edge detection, identifying the edge position of the copper foil 160, and then using morphological operations to improve the continuity of the edge.

[0087] The Sobel operator is used to identify the edge of the copper foil 160 in the image, and two 3x3 convolution kernels can be used to calculate the gradient information of the image in the horizontal direction and the vertical direction respectively, so as to detect the edge.

[0088] Among them, the convolution kernel gradient information G x Used to detect the horizontal edge of the image, G x Emphasize the grayscale change in the horizontal direction. x By performing a convolution operation with the image, the gradient information of the image in the horizontal direction can be obtained.

[0089] Convolution kernel gradient information G y Used to detect the vertical edge of the image, G y Emphasize the grayscale change in the vertical direction. y By performing convolution operation with the image, the gradient information of the image in the vertical direction can be obtained.

[0090] According to the gradient information in the horizontal and vertical directions of the image, the image gradient size and gradient direction are obtained to detect the edge.

[0091] The image gradient size calculation formula is:

[0092]

[0093] The image gradient direction calculation formula is:

[0094]

[0095] Among them, G x is the gradient information in the horizontal direction, G y It is the vertical gradient information.

[0096] Morphological operations can be performed on the edge detection results to fill the small holes between the edges and improve the continuity of the edges.

[0097] Based on the edge detection result, the actual distance can be obtained by calculating the pixel distance between the edges of the copper foil 160 on both sides in the image, and then the back-end analysis server corrects the calculation result to ensure accuracy.

[0098] The pixel distance calculation formula is:

[0099] Width(in pixels)=|x2-x1| (6)

[0100] Among them, (x1, y1) and (x2, y2) are the detection results of the boundary lines on both sides of the copper foil 160, and the width directions x1 and x2 are taken for calculation.

[0101] The conversion formula between actual distance and pixel distance is:

[0102] W = Width (in pixels) × d (7)

[0103] Where W is the actual distance and d is the conversion factor from pixel distance to actual distance.

[0104] The detection results are transmitted to the display screen 140 for real-time display, which is convenient for the operator to monitor; the display screen 140 can display real-time images and width values ​​to provide intuitive detection results; when the detected width exceeds the preset range, the system will automatically send out an alarm signal through the intermediate relay to control the alarm light to remind the operator to make adjustments.

[0105] Furthermore, in order to facilitate subsequent analysis and quality control, data storage and analysis can be performed, and all test data can be stored in the server. The data and results of each test can be stored in association with the corresponding production batch information, which is convenient for subsequent quality tracing and problem analysis. The quality traceability function can help enterprises achieve comprehensive monitoring and management of the production process, improve production efficiency and product quality.

[0106] Furthermore, in the process of transmitting image data, in order to ensure the stability and real-time performance of data transmission, high-speed Ethernet or optical fiber transmission may be used.

[0107] Furthermore, before the entire device is used, the device may be calibrated first, and the calibration process may take into account the distortion parameters and shooting angle of the camera to improve the measurement accuracy.

[0108] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

[0109] In the present invention, unless otherwise clearly specified and limited, a first feature being "above" or "below" a second feature may mean that the first and second features are in direct contact, or the first and second features are in indirect contact through an intermediate medium. Moreover, a first feature being "above", "above" or "above" a second feature may mean that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is higher in level than the second feature. A first feature being "below", "below" or "below" a second feature may mean that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is lower in level than the second feature.

Claims

1. A copper foil width detection device based on visual recognition algorithm, characterized in that: include: Two cameras (110) are respectively located on two sides of the copper foil (160) to be inspected, fixed on the bridge (120), and collect image data of the edges of the copper foil (160) on both sides; A driving device (130) for conveying the copper foil (160) to ensure that the edges on both sides of the copper foil (160) move smoothly within the field of view of the camera (110); The back-end analysis server (150) receives the image data from the camera (110), performs image data preprocessing, calculates the width of the copper foil (160) through a visual recognition algorithm, and transmits the result to the display screen (140) for display.

2. The copper foil width detection device based on visual recognition algorithm according to claim 1 is characterized in that: Also includes: The alarm device (180) emits an audible and visual alarm when the width of the copper foil (160) calculated by the back-end analysis server (150) exceeds a preset range.

3. A copper foil width detection method based on a visual recognition algorithm applied to the device described in claim 1 or 2, characterized in that: The method comprises the steps of: S1, keeping the edges of both sides of the copper foil (160) moving smoothly within the field of view of the two cameras (110); S2, using two cameras (110) to respectively collect image data of the edges of the copper foil (160) on both sides, and sending the image data to a back-end analysis server (150); S3, using the backend analysis server (150) to perform image data preprocessing on the collected image data; S4, calculating the width of the copper foil (160) by a visual recognition algorithm, and transmitting the result to a display screen (140) for display; S5. When the calculated width of the copper foil (160) exceeds a preset range, an audible and visual alarm is issued.

4. The copper foil width detection method based on visual recognition algorithm according to claim 3 is characterized in that: The preprocessing of the image data in S3 includes: graying, denoising and image enhancement preprocessing of the image data.

5. The copper foil width detection method based on visual recognition algorithm according to claim 4 is characterized in that: The image data is preprocessed into grayscale, including converting the captured color image into a grayscale image, and the formula is: Gray=0.299×R+0.587×G+0.114×B (1) Where R, G, and B are the red, green, and blue channels of the image, respectively.

6. The copper foil width detection method based on visual recognition algorithm according to claim 4 is characterized in that: The image data is pre-processed for denoising, including smoothing the image using a Gaussian filter to reduce noise interference. The calculation formula is: Here, σ is the standard deviation of the Gaussian function.

7. The copper foil width detection method based on visual recognition algorithm according to claim 4 is characterized in that: The image data is subjected to image enhancement preprocessing, including using adaptive histogram equalization to improve the contrast of the image and enhance the local detail features of the image, the steps are as follows: S31, dividing the image into several small areas, each of which has a size of M×N, and independently performing histogram equalization on each small area. The conversion function formula of the histogram equalization is: Where I is the pixel value of the small area, H(j) is the histogram, L is the number of gray levels, and M×N is the number of pixels in the small area; S32, then for any bin in the histogram, if its frequency exceeds the threshold, the excess part is evenly distributed to all other bins; S33, performing bilinear interpolation and merging on the equalization results of adjacent small blocks to obtain a final enhanced image.

8. The copper foil width detection method based on visual recognition algorithm according to claim 3 is characterized in that: The visual recognition algorithm uses the Sobel operator to identify the edge of an object in an image, performs image edge detection, identifies the edge position of the copper foil (160), and uses morphological operations to improve the continuity of the edge.

9. The copper foil width detection method based on visual recognition algorithm according to claim 8, characterized in that: The method of using the Sobel operator to identify the edge of the copper foil (160) in the image comprises the steps of: Use two convolution kernels to calculate the gradient information of the image in the horizontal and vertical directions respectively; Among them, the convolution kernel gradient information G x Used to detect the horizontal edge of the image, G x Emphasize the grayscale change in the horizontal direction. x By performing convolution operation with the image, the gradient information of the image in the horizontal direction can be obtained; Convolution kernel gradient information G y Used to detect the vertical edge of the image, G y Emphasize the grayscale change in the vertical direction. y By performing convolution operation with the image, the gradient information of the image in the vertical direction can be obtained; According to the gradient information in the horizontal and vertical directions of the image, the image gradient size and gradient direction are obtained to detect the edge, where: The image gradient size calculation formula is: The image gradient direction calculation formula is: Among them, G x is the gradient information in the horizontal direction, G y It is the vertical gradient information.

10. The copper foil width detection method based on visual recognition algorithm according to claim 9, characterized in that: The calculating of the width of the copper foil (160) includes obtaining an actual distance by calculating the pixel distance between the edges on both sides of the copper foil (160) in the image based on the edge detection result, wherein: The pixel distance calculation formula is: Width(in pixels)=|x2-x1| (6) Wherein, x1 and x2 are the boundary line detection results on both sides of the copper foil (160); The conversion formula between actual distance and pixel distance is: W = Width (in pixels) × d (7) Where W is the actual distance and d is the conversion factor from pixel distance to actual distance.

Images