A battery electrode die cut part defect detection method, apparatus and device

Abstract

This invention discloses a method, apparatus, and device for detecting defects in battery electrode die-cut parts. The method includes: acquiring an image of the die-cut part to be inspected, and extracting the contour information and surface defect point information of the die-cut part from the image; comparing the contour information with a standard contour, and comparing the surface defect point information with a standard defect threshold; if the contour difference between the contour information and the standard contour is greater than a preset threshold, or if the surface defect point information is greater than the standard defect threshold, then the die-cut part to be inspected is determined to be a defective part. The technical solution provided by this invention further improves the identification effect of defects in battery electrode die-cut parts.

Description

Technical Field

[0001] This invention relates to the field of computer vision, and specifically to a method, apparatus, and equipment for detecting defects in die-cut battery electrode parts. Background Technology

[0002] With the rapid development of the new energy industry, the demand for power batteries is increasing daily. The quality of power batteries not only affects battery life but also driving safety. Figure 1 The diagram shows a structural schematic of a square battery cell. Die-cutting is a crucial step in the production of stacked batteries, and the quality of the die-cut parts directly determines the quality of the battery. Traditional defect detection methods based on manual offline sampling suffer from high costs, low detection efficiency, insufficient inspection rate, and the potential to introduce additional damage. Therefore, a device has been developed that uses cameras to take pictures on the production line and then analyzes the defects in the die-cut parts based on these images. However, current technologies for analyzing die-cut part images largely rely on manual judgment. Therefore, it is imperative to further improve the identification of defects in battery electrode die-cut parts. Summary of the Invention

[0003] In view of this, embodiments of the present invention provide a method, apparatus and equipment for detecting defects in battery electrode die-cut parts, thereby improving the identification effect of defects in battery electrode die-cut parts.

[0004] According to a first aspect, the present invention provides a method for detecting defects in battery electrode die-cut parts. The method includes: acquiring a die-cut part image of a current die-cut part to be detected, and extracting contour information and surface defect point information of the die-cut part from the die-cut part image; comparing the contour information with a standard contour, and comparing the surface defect point information with a standard defect threshold; if the contour difference between the contour information and the standard contour is greater than a preset threshold, or if the surface defect point information is greater than the standard defect threshold, then determining that the current die-cut part to be detected is a defective part.

[0005] Optionally, extracting the contour information of the die-cut part from the die-cut part image includes: extracting sub-pixel information of the die-cut part image; performing primitive segmentation on the sub-pixel information to obtain contour point information of the die-cut part; removing error points in the contour point information whose dispersion exceeds a preset threshold, and performing primitive fitting on the remaining contour point information to obtain the contour information.

[0006] Optionally, extracting surface defect point information of the die-cut part from the die-cut part image includes: creating a standard template image and matching the pixels of the die-cut part image with the corresponding pixels in the standard template image; converting the die-cut part image into a binary image based on the successfully matched pixels and the unmatched pixels, and using the binary image as the surface defect point information.

[0007] Optionally, creating the standard template image includes: selecting a preset number of die-cut parts images corresponding to the die-cut parts to be detected in the detection task and aligning and overlaying them; using the average value of each pixel in the overlay image as the pixel value at the corresponding position to generate the standard template image.

[0008] Optionally, matching the pixels of the die-cut part image with the corresponding pixels in the standard template image includes: calculating the standard deviation of the pixels in the overlay image, and calculating a preset standard deviation threshold based on the standard deviation; extracting pixels within a preset neighborhood corresponding to each pixel in the die-cut part image as the center; comparing the pixels extracted in the current neighborhood with the pixels within the corresponding range of the standard template image; counting the number of pixels whose comparison deviation is less than the preset standard deviation threshold; if the number of pixels is greater than a preset number threshold, then the matching of the center pixel corresponding to the current neighborhood is determined to be successful; if the number of pixels is not greater than the preset number threshold, then the matching of the center pixel corresponding to the current neighborhood is determined to be unsuccessful.

[0009] Optionally, converting the die-cut part image into a binary image based on successfully matched pixels and unmatched pixels includes: setting successfully matched pixels in the die-cut part image as first color points and unmatched pixels as second color points to obtain the binary image; and marking connected regions in the binary image based on the connectivity of the first color points and the second color points.

[0010] Optionally, comparing the surface defect point information with a standard defect threshold includes: comparing the number of connected region markers corresponding to the second color points in the binary image and the range of connected markers corresponding to each connected region with the defect number threshold and defect range threshold in the standard defect threshold, respectively; if the surface defect point information is greater than the standard defect threshold, then determining the die-cut part as a defective part includes: if the number of connected region markers is greater than the defect number threshold, or the range of connected markers corresponding to any connected region is greater than the defect range threshold, then determining the current die-cut part to be detected as a defective part.

[0011] According to a second aspect, embodiments of the present invention provide a defect detection device for battery electrode die-cut parts. The device includes: a data extraction unit, configured to acquire a die-cut part image of the current die-cut part to be detected, and extract contour information and surface defect point information of the die-cut part from the die-cut part image; a data comparison unit, configured to compare the contour information with a standard contour, and compare the surface defect point information with a standard defect threshold; and a defect determination unit, configured to determine that the current die-cut part to be detected is a defective part if the contour difference between the contour information and the standard contour is greater than a preset threshold, or if the surface defect point information is greater than the standard defect threshold.

[0012] According to a third aspect, embodiments of the present invention provide a defect detection device for battery electrode die-cut parts. The device includes: an image acquisition sub-device and an image processing sub-device; the image processing sub-device stores computer instructions, and by executing the computer instructions, executes the method provided in any optional embodiment of the first aspect; the image acquisition sub-device includes a light-transmitting conveyor belt, a line scan camera, a first line scan light source, a second line scan light source, a first fiber optic sensor, a second fiber optic sensor, and a suction cup sorting device; the line scan camera is located above the light-transmitting conveyor belt, and the first line scan light source and the second line scan light source are respectively located above and below the light-transmitting conveyor belt. The illumination direction of the light source and the second line scan light source is towards the image capture position of the line scan camera. The first fiber optic sensor and the second fiber optic sensor are disposed opposite each other above and below the light-transmitting conveyor belt. The first fiber optic sensor is disposed adjacent to the line scan camera and close to the incoming direction of the light-transmitting conveyor belt. The suction cup sorting device includes a first suction cup arm and a second suction cup arm, which are respectively located at both ends of the transmission direction of the light-transmitting conveyor belt. The line scan camera, the first line scan light source, the second line scan light source, the first fiber optic sensor, the second fiber optic sensor, and the suction cup sorting device are all communicatively connected to the image processing sub-device.

[0013] According to a fourth aspect, embodiments of the present invention provide a computer-readable storage medium storing computer instructions for causing the computer to perform the method described in the first aspect, or any alternative embodiment of the first aspect.

[0014] The technical solution provided in this application has the following advantages:

[0015] The technical solution provided in this application, based on the acquired image of the die-cut part to be inspected, simultaneously extracts the contour information and surface defect point information of the die-cut part from the image; and compares the contour information with a standard contour and the surface defect point information with a standard defect threshold, respectively; if the contour information differs significantly from the standard contour, or the surface defect point information differs significantly from the standard defect threshold, the die-cut part to be inspected is determined to be defective. This improves the accuracy of defect identification in the die-cut part to be inspected.

[0016] Furthermore, particularly in the extraction of die-cut part contour information, this embodiment of the invention first selects multiple die-cut parts to be inspected and overlays them to create a standard template image of the die-cut parts that represents the average level of the die-cut parts produced this time. This avoids the problem of unstable and inaccurate defect detection accuracy caused by the defect comparison standard being too high or too low. Then, the die-cut part image of the die-cut part to be inspected and the standard template image are matched for each pixel, and the pixels that match successfully and those that fail to match are set to different colors. Then, the current die-cut part image is judged based on the standard of the number and range of pixels that fail to match. On the one hand, this makes the defects of the die-cut parts have a certain degree of acceptability, avoiding the defect detection standard being too harsh, which would cause die-cut parts that do not affect normal use to be classified as defective parts, thereby avoiding waste of resources. On the other hand, it analyzes the surface defects of the die-cut parts from two perspectives, improving the accuracy of die-cut part defect detection. Attached Figure Description

[0017] The features and advantages of the invention will be more clearly understood by referring to the accompanying drawings, which are schematic and should not be construed as limiting the invention in any way. In the drawings:

[0018] Figure 1 A schematic diagram of a square battery cell is shown.

[0019] Figure 2 This diagram illustrates the steps of a method for detecting defects in die-cut battery electrodes according to one embodiment of the present invention.

[0020] Figure 3 A flowchart illustrating a method for detecting defects in die-cut battery electrodes according to one embodiment of the present invention is shown.

[0021] Figure 4 A schematic diagram of a die-cut part according to one embodiment of the present invention is shown;

[0022] Figure 5 This image shows a schematic diagram of a die-cut part containing defects according to one embodiment of the present invention;

[0023] Figure 6 This diagram illustrates the structure of a defect detection device for battery electrode die-cut parts according to one embodiment of the present invention.

[0024] Figure 7 This diagram illustrates the structure of a defect detection device for battery electrode die-cut parts according to one embodiment of the present invention.

[0025] Figure 8 The diagram illustrates the workflow of a battery electrode die-cutting defect detection device according to one embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] Please see Figure 2 and Figure 3 In one embodiment, a method for detecting defects in battery electrode die-cut parts specifically includes the following steps:

[0028] Step S101: Obtain the die-cut image of the current die-cut part to be inspected, and extract the contour information and surface defect point information of the die-cut part from the die-cut image.

[0029] Step S102: Compare the contour information with the standard contour, and compare the surface defect point information with the standard defect threshold.

[0030] Step S103: If the contour difference between the contour information and the standard contour is greater than a preset threshold, or if the surface defect point information is greater than the standard defect threshold, then the current die-cut part to be inspected is determined to be a defective part.

[0031] Specifically, in this embodiment, an image of the die-cut part is first acquired using an image acquisition device, such as... Figure 4As shown, contour extraction and image feature extraction algorithms are then used to extract contour information and surface defect point information from the die-cut part image, respectively. Contour information refers not only to the contour around the electrode sheet but also to the contour of the electrode sheet surface, such as the uneven structure of the electrode sheet surface. The contour information is then compared with a complete standard contour to determine whether the die-cut part includes defects such as broken contour edges. Specifically, a deviation threshold or standard deviation threshold can be used to measure the deviation between contours. Furthermore, the surface defect point information is compared with a standard defect threshold representing a flat and smooth surface of the die-cut part to determine whether the die-cut part includes defects such as foil leakage, black and white spots, scratches, cracks, and uneven points. If any aspect of the contour information or surface defect information does not meet the preset standard, the die-cut part corresponding to the current image is identified as a defective part, thus accurately identifying various defects in the die-cut part and improving the accuracy of defect identification.

[0032] Specifically, in one embodiment, step S101 above includes the following steps:

[0033] Step 1: Extract subpixel information from the die-cut part image.

[0034] Step 2: Perform primitive segmentation on the subpixel information to obtain the contour point information of the die-cut part.

[0035] Step 3: Remove error points in the contour point information whose dispersion exceeds the preset threshold, and perform primitive fitting on the remaining contour point information to obtain contour information.

[0036] Specifically, in this embodiment, sub-pixel information of the die-cut part image is first extracted to enhance its resolution. The extraction method is as follows: first, the gradient map of the die-cut part image is obtained using the Sobel operator; then, a quadratic polynomial interpolation operation is performed on the gradient map to obtain the sub-pixel information of the die-cut part image. Next, Hough transform is performed by defining a detection window to segment primitives, obtaining contour points and partial contour lines in the die-cut part image. Then, a primitive fitting process is performed on the contour point information to obtain a clear contour of the die-cut part. Specifically, the points in the contour point information are first fitted with a straight line using the least squares method. After obtaining the fitted straight line, contour points that are too far from the fitted straight line are removed. Then, the contour is fitted based on the remaining contour points to obtain accurate contour information.

[0037] Specifically, in one embodiment, step S101 above further includes the following step:

[0038] Step 4: Create a standard template image and match the pixels of the die-cut part image with the corresponding pixels in the standard template image.

[0039] Step 5: Convert the die-cut part image into a binary image based on the successfully matched pixels and the unmatched pixels, and use the binary image as surface defect point information.

[0040] Specifically, in this embodiment, the method for extracting surface defect point information of the die-cut part is as follows: First, a standard image template is created. Then, the image of the die-cut part to be inspected is matched pixel by pixel with the standard template image. If the pixel difference is large, causing the matching to fail, the pixels that failed to match are set to a different color than the pixels that matched successfully, thereby distinguishing them and obtaining accurate surface defect point information of the die-cut part. The image of the die-cut part containing the defect points can be referenced. Figure 5 .

[0041] Specifically, in this embodiment, the steps for creating a standard template image include:

[0042] 1. Select the pre-set number of die-cut parts images to be inspected in the inspection task and align and overlay them.

[0043] 2. Use the average value of each pixel in the overlay image as the pixel value at the corresponding position to generate a standard template image.

[0044] Specifically, by selecting a preset number of die-cut parts images from the inspection task and aligning and overlaying them, in this embodiment, the first 30 die-cut parts are selected. Then, the average value of the pixel value corresponding to each pixel position in the overlay image is calculated. For example, if 30 pixel values ​​are overlaid at the first pixel position, then the pixel value after overlay at this position is the average of 30 pixel values. This is used to create a standard template image, thereby ensuring that the color in the template can represent the average level of the die-cut parts produced in this batch. This avoids the problem of being too strict or too lenient when specifying color standards, thus avoiding both die-cut parts that meet the usage conditions being misjudged as defective and defective die-cut parts being misjudged as normal.

[0045] Specifically, in one embodiment, step four above includes the following steps:

[0046] Step 6: Calculate the standard deviation of the pixels in the overlay image, and calculate the preset standard deviation threshold based on the standard deviation.

[0047] Step 7: Using each pixel in the die-cut part image as the center, extract the pixels within the preset neighborhood corresponding to each center.

[0048] Step 8: Compare the pixels extracted from the current neighborhood with the pixels within the corresponding range of the standard template image.

[0049] Step 9: Count the number of pixels whose comparison deviation is less than the preset standard deviation threshold.

[0050] Step 10: If the number of pixels is greater than the preset threshold, then the center pixel of the current neighborhood is considered to have been successfully matched.

[0051] Step 11: If the number of pixels is not greater than the preset threshold, then the matching of the center pixel corresponding to the current neighborhood is determined to be unsuccessful.

[0052] Specifically, in this embodiment, the standard deviation of multiple pixel values ​​at each pixel location in the overlay image is first calculated. For example, the first pixel location in the overlay image has 30 pixel values, and the standard deviation is calculated as the standard deviation of these 30 pixel values. Similarly, the standard deviation of each pixel location is obtained. Then, the standard deviation threshold used to determine whether each pixel is successfully matched is calculated using the standard deviation of each pixel location. In this embodiment, the standard deviation threshold is plus or minus three times the standard deviation. Referring to the normal distribution theory, if the deviation between any pixel value of the die-cut part to be inspected and the corresponding pixel value in the standard template image falls within the above-mentioned range of plus or minus three times the standard deviation, it can be determined that the pixel value of the current pixel is not significantly different from the standard pixel value and is not a defect point.

[0053] Furthermore, in this embodiment, to further improve the accuracy of pixel detection in the die-cut part image, pixels within a preset neighborhood (e.g., within a 15*15 neighborhood) corresponding to each pixel in the die-cut part image are extracted. For any neighborhood, all extracted pixels in the current neighborhood are compared with pixels within the corresponding range of the standard template image, and the aforementioned standard deviation determination step is performed. Then, the number of pixels in the current neighborhood with a comparison deviation less than a preset standard deviation threshold is counted, i.e., the number of pixels whose pixel values ​​are not significantly different from the standard pixel values. Only when the number exceeds a preset threshold (e.g., the preset threshold is 70% of the total number of pixels in the current range) is the center pixel corresponding to the current neighborhood considered a normal point; otherwise, it is a defective point. This further improves the accuracy of pixel detection in the die-cut part image.

[0054] Let's explain with a specific example:

[0055] There are three types of images: an overlay image of the die-cut part to be detected, a standard template image, and a generated standard template image. Taking a pixel at the center of the die-cut part image (hereinafter referred to as the center pixel) as an example, a 15*15 neighborhood is cropped from this pixel in the die-cut part image, resulting in all pixels within this 15*15 area. Then, each pixel within this area is compared with its corresponding pixel in the standard template image to calculate the deviation, thus obtaining multiple deviation values. For each pixel within this area, multiple positive and negative three-times standard deviations can be calculated at its corresponding position in the overlay image, resulting in multiple standard deviation thresholds. Then, each deviation value is compared according to its positional correspondence with the standard deviation thresholds, and the number of pixels with deviation values ​​less than the standard deviation thresholds is counted (hereinafter referred to as the number of pixels satisfying the condition). Finally, the ratio of the number of pixels satisfying the condition to the total number of pixels in the current neighborhood is calculated. If the calculated ratio exceeds 70%, the current center pixel is considered a successfully matched pixel.

[0056] Specifically, in one embodiment, step five above includes the following steps:

[0057] Step 12: Set the successfully matched pixels in the die-cut part image to the first color point, and set the unmatched pixels to the second color point to obtain a binary image.

[0058] Step 13: Based on the connectivity of the first and second color points, mark the connected regions in the binary image.

[0059] Specifically, in this embodiment, the successfully matched pixels and the unmatched pixels are set to different colors, for example, successfully matched pixels are set to black and unmatched pixels are set to white. Then, connected regions are marked according to the connectivity of pixels with the same color, making the defects of the die-cut parts more obvious, thereby further improving the accuracy of defect identification of the die-cut parts.

[0060] Specifically, in one embodiment, step S102 above includes the following steps:

[0061] Step Fourteen: Compare the number of connected region markers corresponding to the second color points in the binary image and the range of connected region markers corresponding to each connected region with the defect number threshold and defect range threshold in the standard defect threshold.

[0062] Based on step fourteen, step S103 above specifically includes the following steps:

[0063] Step 15: If the number of connected region markers is greater than the defect number threshold, or the range of connected markers corresponding to any connected region is greater than the defect range threshold, then the current die-cut part to be inspected is determined to be a defective part.

[0064] Specifically, this embodiment divides the determination of defective parts into contour determination and surface defect determination. If either aspect fails to meet the preset conditions, the die-cut part to be inspected is determined to be a defective part. Surface defect determination is further divided into defect number determination and defect range determination. Each connected region of the marked second color value (white, i.e., defect location) in the binary image is compared with a defect range threshold. If the area of ​​a connected region is greater than the standard area of ​​the defect range threshold, the current die-cut part is determined to be a defective part. Furthermore, the number of connected regions corresponding to the second color value in the binary image is compared with a defect number threshold. If it exceeds the defect number threshold, the current die-cut part is considered to have too many defect points and is also determined to be a defective part. Therefore, only when the die-cut part image simultaneously meets all the above conditions can the corresponding die-cut part be determined as a normal part, thereby further improving the accuracy of die-cut part defect detection.

[0065] Through the above steps, the technical solution provided in this application, for the acquired image of the die-cut part to be inspected, simultaneously extracts the contour information and surface defect point information of the die-cut part from the image; and compares the contour information with a standard contour and the surface defect point information with a standard defect threshold, respectively; if the contour information differs significantly from the standard contour, or the surface defect point information differs significantly from the standard defect threshold, the die-cut part to be inspected is determined to be defective. This improves the accuracy of defect identification in the die-cut part to be inspected.

[0066] Furthermore, particularly in the extraction of die-cut part contour information, this embodiment of the invention first selects multiple die-cut parts to be inspected and overlays them to create a standard template image of the die-cut parts that represents the average level of the die-cut parts produced this time. This avoids the problem of unstable and inaccurate defect detection accuracy caused by the defect comparison standard being too high or too low. Then, the die-cut part image of the die-cut part to be inspected and the standard template image are matched for each pixel, and the pixels that match successfully and those that fail to match are set to different colors. Then, the current die-cut part image is judged based on the standard of the number and range of pixels that fail to match. On the one hand, this makes the defects of the die-cut parts have a certain degree of acceptability, avoiding the defect detection standard being too harsh, which would cause die-cut parts that do not affect normal use to be classified as defective parts, thereby avoiding waste of resources. On the other hand, it analyzes the surface defects of the die-cut parts from two perspectives, improving the accuracy of die-cut part defect detection.

[0067] like Figure 6 As shown, this embodiment also provides a defect detection device for battery electrode die-cut parts, the device comprising:

[0068] The data extraction unit 101 is used to acquire an image of the die-cut part to be inspected, and to extract the contour information and surface defect point information of the die-cut part from the image. For details, please refer to the relevant description of step S101 in the above method embodiment, which will not be repeated here.

[0069] The data comparison unit 102 is used to compare the contour information with the standard contour and to compare the surface defect point information with the standard defect threshold. For details, please refer to the relevant description of step S102 in the above method embodiments, which will not be repeated here.

[0070] The defect determination unit 103 is used to determine that the current die-cut part to be inspected is a defective part if the contour difference between the contour information and the standard contour is greater than a preset threshold, or if the surface defect point information is greater than the standard defect threshold. For details, please refer to the relevant description of step S103 in the above method embodiment, which will not be repeated here.

[0071] The battery electrode die-cutting defect detection device provided in this embodiment of the invention is used to execute the battery electrode die-cutting defect detection method provided in the above embodiment. Its implementation method and principle are the same. For details, please refer to the relevant description of the above method embodiment, which will not be repeated here.

[0072] Through the collaborative efforts of the aforementioned components, the technical solution provided in this application extracts both the contour information and surface defect point information of the die-cut part from the acquired image of the die-cut part to be inspected. The contour information is then compared with a standard contour, and the surface defect point information is compared with a standard defect threshold. If the difference between the contour information and the standard contour is significant, or the difference between the surface defect point information and the standard defect threshold is significant, the die-cut part to be inspected is determined to be defective. This improves the accuracy of defect identification in the die-cut part to be inspected.

[0073] Furthermore, particularly in the extraction of die-cut part contour information, this embodiment of the invention first selects multiple die-cut parts to be inspected and overlays them to create a standard template image of the die-cut parts that represents the average level of the die-cut parts produced this time. This avoids the problem of unstable and inaccurate defect detection accuracy caused by the defect comparison standard being too high or too low. Then, the die-cut part image of the die-cut part to be inspected and the standard template image are matched for each pixel, and the pixels that match successfully and those that fail to match are set to different colors. Then, the current die-cut part image is judged based on the standard of the number and range of pixels that fail to match. On the one hand, this makes the defects of the die-cut parts have a certain degree of acceptability, avoiding the defect detection standard being too harsh, which would cause die-cut parts that do not affect normal use to be classified as defective parts, thereby avoiding waste of resources. On the other hand, it analyzes the surface defects of the die-cut parts from two perspectives, improving the accuracy of die-cut part defect detection.

[0074] Figure 7 An embodiment of the present invention illustrates a defect detection device for battery electrode die-cut parts, the device comprising an image acquisition sub-device 901 and an image processing sub-device 902;

[0075] One or more modules are stored in the image processing sub-device 902. When executed by the image processing sub-device 902, the methods in the above method embodiments are executed.

[0076] Image acquisition sub-device 901 includes a light-transmitting conveyor belt 5, a line scan camera 1, a first line scan light source 2, a second line scan light source 3, a first fiber optic sensor, a second fiber optic sensor, and a suction cup sorting device.

[0077] A line scan camera 1 is positioned above the light-transmitting conveyor belt 5. A first line scan light source 2 and a second line scan light source 3 are positioned above and below the light-transmitting conveyor belt 5, respectively. The illumination direction of the first and second line scan light sources 2 and 3 is towards the image capture position of the line scan camera 1. The first line scan light source 2 is used to illuminate or enhance the contrast of surface defects (leaking foil, black and white spots, scratches, cracks, bumps, and edge damage) on the positive and negative electrode die-cut parts 4 during image capture, making defects easier to extract during image processing. The second line scan light source 3, located below the light-transmitting conveyor belt 5, is used to illuminate the outline of the positive and negative electrode die-cut parts 4 during image capture. Combined with the highly transparent conveyor belt, this allows for the extraction of a clearly contrasting black-and-white outline of the die-cut parts 4 during image processing.

[0078] The first and second fiber optic sensors are positioned opposite each other above and below the light-transmitting conveyor belt 5. The first fiber optic sensor is adjacent to the line scan camera 1 and close to the direction of the incoming part on the light-transmitting conveyor belt 5. Thus, the die-cut part 4 is accurately detected by the first and second fiber optic sensors before it reaches the camera's image capture position. After the first and second fiber optic sensors send detection signals to the image processing sub-device 902, the image processing sub-device 902 promptly turns on the line scan light source and controls the line scan camera 1 to start taking pictures, ensuring that the die-cut part 4 is accurately captured by the line scan camera 1.

[0079] The suction cup sorting device includes a first suction cup arm and a second suction cup arm, which are located at both ends of the light-transmitting conveyor belt 5 in the transmission direction, respectively.

[0080] The line scan camera 1, the first line scan light source 2, the second line scan light source 3, the first fiber optic sensor, the second fiber optic sensor, and the suction cup sorting device are all communicatively connected to the image processing sub-device 902.

[0081] The image processing sub-equipment 902 also includes a PLC controller, which sorts the die-cut parts 4 according to the defect detection results of the die-cut parts 4.

[0082] like Figure 8As shown, the working process of the above-mentioned battery electrode die-cutting defect detection equipment is as follows:

[0083] 1. The positive and negative battery electrode winding material arrives at the cutting station, and the cutting mechanism cuts the material to obtain the cut die-cut part 4;

[0084] 2. The first suction cup arm located at the starting direction of the light-transmitting conveyor belt 5 picks up the die-cut part 4 and places it on the light-transmitting conveyor belt 5, and the light-transmitting conveyor belt 5 carries the die-cut part 4 forward at a constant speed;

[0085] 3. During the movement of the die-cut part 4, the first and second fiber optic sensors on the upper and lower parts of the conveyor belt are triggered, so that the image processing sub-device receives the sensing signal and turns on the first and second line scan light sources in response to the sensing signal. The line scan camera 1 starts to take pictures according to the set image height (number of lines).

[0086] 4. The image processing sub-device acquires the image captured by the line scan camera 1, executes the image processing method in the method embodiment based on the image data, thereby determining whether the die-cut part 4 is a defective part, and outputs an OK / NG signal to the PLC.

[0087] 5. Based on the detection data, the PLC determines whether the corresponding die-cut part 4 is an OK product, thereby determining whether the die-cut part 4 is qualified in this inspection. Then, the qualified products flow to the next station, and the unqualified products are picked up by the second suction cup arm and sent to the waste removal station.

[0088] The battery electrode die-cutting defect detection equipment provided in this invention, through the cooperation of upper and lower linear scanning light sources, can not only easily extract the contour of the die-cutting part, but also provide a high contrast between the die-cutting part and the background for defects such as foil leakage, black and white spots, scratches, cracks, unevenness, and edge damage, facilitating algorithm processing. Furthermore, it can completely replace the original manual offline sampling inspection method for defect detection of lithium battery positive and negative electrode die-cutting parts, reducing inspection costs and solving problems such as low inspection efficiency, insufficient inspection rate, and additional damage during inspection. In a specific implementation scenario, the battery electrode die-cutting defect detection equipment provided in this invention achieves a dimensional defect detection deviation of no more than 0.02 mm, a surface defect detection accuracy of over 99.6%, and can guarantee real-time online detection at 80 meters per minute, meeting the high-speed operation cycle of the production line.

[0089] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The implemented program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium can also include combinations of the above types of memory.

[0090] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for detecting defects in die-cut battery electrode parts, characterized in that, The method includes: Obtain an image of the die-cut part to be inspected, and extract the contour information and surface defect point information of the die-cut part from the image; The contour information is compared with the standard contour, and the surface defect point information is compared with the standard defect threshold. If the contour difference between the contour information and the standard contour is greater than a preset threshold, or if the surface defect point information is greater than the standard defect threshold, then the current die-cut part to be detected is determined to be a defective part. Extracting surface defect point information of the die-cut parts from the die-cut part images includes: selecting a preset number of die-cut part images corresponding to the die-cut parts to be inspected in the inspection task and aligning and overlaying them; using the mean value of each pixel in the overlay image as the pixel value at the corresponding position to generate a standard template image; calculating the standard deviation of the pixels in the overlay image, and calculating a preset standard deviation threshold based on the standard deviation, wherein the preset standard deviation threshold is plus or minus three times the standard deviation of the pixels in the overlay image; taking each pixel in the die-cut part image as the center, extracting the pixels within a preset neighborhood corresponding to each center; and comparing the pixels extracted in the current neighborhood with the pixels within the corresponding range of the standard template image. Yes; count the number of pixels whose comparison deviation is less than the preset standard deviation threshold; if the number of pixels is greater than the preset quantity threshold, then the center pixel corresponding to the current neighborhood is determined to be successfully matched; if the number of pixels is not greater than the preset quantity threshold, then the center pixel corresponding to the current neighborhood is determined to be unmatched; set the successfully matched pixels in the die-cut part image as first color points, and set the unmatched pixels as second color points to obtain a binary image; based on the connectivity of the first color points and the second color points, mark the connected regions in the binary image according to the connectivity of the same color, and use the binary image as the surface defect point information; The step of comparing the surface defect point information with the standard defect threshold includes: comparing the number of connected region markers corresponding to the second color points in the binary image and the range of connected markers corresponding to each connected region with the defect number threshold and defect range threshold in the standard defect threshold, respectively; the step of determining the die-cut part as a defective part if the surface defect point information is greater than the standard defect threshold includes: if the number of connected region markers is greater than the defect number threshold, or the range of connected markers corresponding to any connected region is greater than the defect range threshold, then the current die-cut part to be detected is determined to be a defective part.

2. The method according to claim 1, characterized in that, Extracting the contour information of the die-cut part from the die-cut part image, including: Extract subpixel information from the die-cut part image; The subpixel information is segmented into primitives to obtain the contour point information of the die-cut part; Error points with a dispersion exceeding a preset threshold in the contour point information are removed, and the remaining contour point information is fitted with primitives to obtain the contour information.

3. A defect detection device for battery electrode die-cut parts, characterized in that, The device includes: The data extraction unit is used to acquire the die-cut image of the current die-cut part to be inspected, and extract the contour information and surface defect point information of the die-cut part from the die-cut part image; the extraction of surface defect point information of the die-cut part from the die-cut part image includes: selecting the die-cut part images corresponding to the first preset number of die-cut parts to be inspected in the inspection task and aligning and superimposing them; using the mean value of each pixel in the superimposed image as the pixel value of the corresponding position to generate a standard template image; calculating the standard deviation of the pixels in the superimposed image, and calculating a preset standard deviation threshold based on the standard deviation; taking each pixel in the die-cut part image as the center, extracting the pixels in the preset neighborhood corresponding to each center; and comparing the pixels extracted in the current neighborhood with the standard template image. The pixels within the corresponding range of the board image are compared; the number of pixels with a comparison deviation less than the preset standard deviation threshold is counted; if the number of pixels is greater than the preset number threshold, the center pixel corresponding to the current neighborhood is determined to be successfully matched; if the number of pixels is not greater than the preset number threshold, the center pixel corresponding to the current neighborhood is determined to be unmatched; the successfully matched pixels in the die-cut part image are set as first color points, and the unmatched pixels are set as second color points to obtain a binary image; based on the connectivity of the first color points and the second color points, connected regions are marked in the binary image according to the connectivity of the same color, and the binary image is used as the surface defect point information; A data comparison unit is used to compare the contour information with a standard contour and to compare the surface defect point information with a standard defect threshold. The comparison of the surface defect point information with the standard defect threshold includes: comparing the number of connected region markers corresponding to the second color points in the binary image and the range of connected markers corresponding to each connected region with the defect number threshold and defect range threshold in the standard defect threshold, respectively. If the surface defect point information is greater than the standard defect threshold, the die-cut part is determined to be a defective part, which includes: if the number of connected region markers is greater than the defect number threshold, or the range of connected markers corresponding to any connected region is greater than the defect range threshold, then the current die-cut part to be detected is determined to be a defective part. The defect determination unit is used to determine that the current die-cut part to be inspected is a defective part if the contour difference between the contour information and the standard contour is greater than a preset threshold, or if the surface defect point information is greater than the standard defect threshold.

4. A defect detection device for battery electrode die-cut parts, characterized in that, The device includes: an image acquisition sub-device and an image processing sub-device; The image processing sub-device stores computer instructions, and by executing the computer instructions, the method described in any one of claims 1-2 is performed; The image acquisition sub-device includes a light-transmitting conveyor belt, a line scan camera, a first line scan light source, a second line scan light source, a first fiber optic sensor, a second fiber optic sensor, and a suction cup sorting device. The line scan camera is located above the light-transmitting conveyor belt. The first line scan light source and the second line scan light source are located above and below the light-transmitting conveyor belt, respectively. The illumination direction of the first line scan light source and the second line scan light source is towards the image capture position of the line scan camera. The first fiber optic sensor and the second fiber optic sensor are arranged opposite each other above and below the light-transmitting conveyor belt. The first fiber optic sensor is arranged adjacent to the line scan camera and close to the incoming direction of the light-transmitting conveyor belt. The suction cup sorting device includes a first suction cup arm and a second suction cup arm, which are located at opposite ends of the light-transmitting conveyor belt in the transmission direction. The line scan camera, the first line scan light source, the second line scan light source, the first fiber optic sensor, the second fiber optic sensor, and the suction cup sorting device are all communicatively connected to the image processing sub-device.

5. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to perform the method as described in any one of claims 1-2.

Images