Defect detection method, system, device, equipment and medium for pole piece

Abstract

This application discloses a method, system, apparatus, device, and medium for detecting defects in electrode sheets. The method includes acquiring a test image, which includes the striped edges of the electrode sheet and an electrode sheet region of a preset width based on the striped edges. The test image is obtained by cropping an initial image of the electrode sheet based on the striped edges. The method further includes detecting the striped edges and target edges of a target region within the electrode sheet region, and determining a step detection result based on the distance between the striped edges and the target edges. It also includes performing aluminum leakage detection on the electrode sheet region in the test image to obtain an aluminum leakage detection result. If the step detection result indicates that the electrode sheet region includes a step, and / or the aluminum leakage detection result indicates that the electrode sheet region has aluminum leakage, then the method determines that the electrode sheet has a defect. Thus, compared to the manual sampling method in related technologies, the electrode sheet defect detection method in this application not only improves detection accuracy but also increases detection efficiency.

Description

Technical Field

[0001] This application relates to the field of battery technology, and in particular to a method, system, device, equipment and medium for detecting defects in electrode sheets. Background Technology

[0002] In the manufacturing process of power batteries, the processing quality of the electrode sheets directly affects the safety and reliability of the battery. Electrode sheet slitting is one of the key processes, the purpose of which is to cut the coated electrode sheets to the specified dimensions. However, during the slitting process, due to the action of the cutting tools, defects such as planar steps or exposed current collectors (aluminum foil) are easily generated at the edges of the electrode sheets. These defects may cause serious problems such as battery short circuits or even fires. Therefore, effective identification and control of electrode sheet edge defects is of great significance.

[0003] In related technologies, manual sampling is typically used to detect defects at the edge of the electrode sheets. This involves observing the cut samples under a microscope to determine if there are steps or aluminum leakage. This method relies on the operator's experience and judgment, resulting in low detection efficiency, poor consistency, and difficulty in meeting the full inspection requirements of large-scale production. Furthermore, it carries a high risk of missed detections and high labor costs. Summary of the Invention

[0004] In view of this, embodiments of this application provide at least one method, system, apparatus, device, and medium for detecting defects in electrode sheets.

[0005] The technical solution of this application embodiment is implemented as follows:

[0006] In a first aspect, embodiments of this application provide a method for detecting defects in an electrode sheet, the method comprising:

[0007] The image to be tested is obtained. The image to be tested includes the strip edges of the electrode and the electrode area with a preset width based on the strip edges. The image to be tested is obtained by cropping the initial image of the electrode based on the strip edges.

[0008] The process detects the slit edges and target edges of the target region within the electrode area, and determines the step detection result based on the distance between the slit edges and the target edges. The step detection result is used to characterize whether the electrode area includes steps. A step is a stepped structure formed on the edge of the electrode due to slit processing. The target edge is any edge on the electrode other than the slit edges.

[0009] Aluminum leakage detection is performed on the electrode region in the image to be tested to obtain the aluminum leakage detection result; the aluminum leakage detection result is used to characterize whether aluminum leakage occurs in the electrode region.

[0010] If the step detection results indicate that the electrode area includes steps, and / or the aluminum leakage detection results indicate that the electrode area has aluminum leakage, then the electrode is determined to have a defect.

[0011] In this embodiment, the image to be tested is first obtained by cropping the edges to remove redundant data, thereby reducing interference from irrelevant areas and improving detection efficiency. Secondly, by combining step detection and aluminum leakage detection methods, the presence of steps or aluminum leakage is determined, achieving multi-dimensional detection of electrode quality. Finally, if any detection result is abnormal, a defect in the electrode is determined. Compared to the manual sampling method in related technologies, the electrode identification method provided in this embodiment not only improves detection accuracy but also increases detection efficiency.

[0012] In some embodiments, aluminum leakage detection is performed on the electrode region in the image to be tested to obtain an aluminum leakage detection result, including: using a trained aluminum leakage detection model to perform a first aluminum leakage detection on the electrode region in the image to be tested to obtain a first detection result; reconstructing the image to be tested and determining the reconstruction error between the reconstructed image and the image to be tested; if the first detection result indicates that aluminum leakage exists in the electrode region and / or the reconstruction error does not meet the conditions, determining that aluminum leakage detection result indicates that aluminum leakage exists in the electrode region.

[0013] In this embodiment, a trained aluminum leakage detection model is used to detect aluminum leakage in the image under test, improving the detection accuracy. Simultaneously, reconstruction processing is combined with aluminum leakage detection on the image under test, reducing the possibility of missed detections. Compared to a single detection method, this embodiment improves the robustness of aluminum leakage detection through multi-criteria fusion.

[0014] In some embodiments, the image to be tested is reconstructed, and the reconstruction error between the reconstructed image and the image to be tested is determined, including: performing dimensional compression processing on the image to be tested to obtain a multi-dimensional feature vector; performing reconstruction processing on the multi-dimensional feature vector to obtain a reconstructed image; and determining the mean square error of the reconstructed image relative to the image to be tested as the reconstruction error.

[0015] In this embodiment, a method for extracting features through dimensionality reduction and reconstructing the image is used to extract the core features of the image. Mean squared error is then used to measure the image reconstruction quality, thereby effectively distinguishing between normal and abnormal images. This can compensate for missed detections caused by using a trained aluminum leakage detection model, thus improving the accuracy of aluminum leakage detection.

[0016] In some embodiments, detecting the strip edges and target edges of the target region in the electrode region, and determining the step detection result based on the distance between the strip edges and the target edges, includes: separating the electrode body region from the target region in the electrode region based on a preset grayscale threshold to obtain the target region; detecting the strip edges and target edges in the target region to obtain the position information of multiple first pixels of the strip edges and the position information of multiple second pixels of the target edges; the pixel width of the strip edges and target edges in the step region is a single pixel; determining the distance between the strip edges and the target edges based on the position information of the multiple first pixels and the position information of the multiple second pixels, and determining the step detection result based on the distance between the strip edges and the target edges.

[0017] In this embodiment, the separation of the electrode body area and the target area is achieved by setting a grayscale threshold, which helps to accurately locate the target area where there may be steps; then, the precise position information of the strip edge and the target edge is obtained by pixel-level edge detection, thereby providing a reliable basis for determining the width information of the target area.

[0018] In some embodiments, determining a step detection result based on the position information of multiple first pixels and multiple second pixels includes: determining multiple pixel pairs, each including a first pixel and a second pixel, based on the position information of multiple first pixels and multiple second pixels; the first pixel and the second pixel in each pixel pair are in the same pixel column or the same pixel row; for each pixel pair, determining the pixel distance between the first pixel and the second pixel based on the position information of the first pixel and the second pixel in the pixel pair; determining the maximum pixel distance among the pixel distances corresponding to the multiple pixel pairs as the width information of the target region in the electrode area; and determining the step detection result as the electrode area including a step if the width information of the target region is greater than or equal to a preset width.

[0019] In this embodiment, the width of the step can be quantified through pixel pair matching and pixel distance calculation, thereby achieving automatic measurement of the step size. Compared with traditional manual visual inspection, this method not only improves the detection accuracy but also provides real-time feedback of the detection results, offering data support for process adjustments.

[0020] In some embodiments, the method further includes: acquiring an initial image of the electrode sheet; the electrode sheet is a single-sided electrode sheet obtained by slicing an initial electrode sheet with two sides; detecting the slicing edges of the initial image; and cropping the detected initial image based on the slicing edges and the cropping edges whose distance to the slicing edges is within a preset distance to obtain the image to be tested.

[0021] In this embodiment, an initial image of the initial electrode sheet of both sides of the electrode is obtained after striping. Then, a striping edge detection operation is performed, and a cutting edge is determined based on a preset distance. The initial image is then cropped to generate a more accurate image to be tested. This method effectively reduces the detection range, improves the accuracy of defect identification, and thus enhances overall detection efficiency.

[0022] In some embodiments, the above method is applied to the die-cutting and slitting process; the initial electrode sheet after slitting includes two electrode sheets with two single-sided tabs; the method further includes: for each electrode sheet with a single-sided tab, acquiring two initial images obtained by acquiring the slitting edges of the electrode sheet using two image acquisition devices respectively; and performing slitting edge detection and cropping on each initial image to obtain the image to be tested.

[0023] In this embodiment, by employing two image acquisition devices to initially acquire images of the upper and lower sides of the electrode sheet on one side of the electrode tab during the die-cutting and slitting process, and combining this with the operation of slitting edge detection and trimming, efficient and accurate detection of slitting edge defects in the electrode sheet on one side of the electrode tab is achieved. This operation method ensures that both the upper and lower sides of the electrode sheet on one side of the electrode tab can be fully detected, thereby avoiding missed detections and improving the comprehensiveness of the electrode sheet inspection.

[0024] In some embodiments, the method further includes: determining that the defect detection result of the electrode sheet is unqualified when the step detection result indicates that the electrode sheet area includes a step, and / or the aluminum leakage detection result indicates that the electrode sheet area has aluminum leakage; and sending a prompt message to the slitting equipment indicating that the cutting tool of the slitting equipment is abnormal when there are multiple consecutive electrode sheets with unqualified defect detection results.

[0025] In this embodiment of the application, by setting up continuous NG judgment logic, the equipment status can be fed back in a timely manner when continuous defects occur, which helps operators to quickly find the cause of the fault, prevent the generation of batch defective products, and improve production stability.

[0026] Secondly, embodiments of this application provide a defect detection system for electrode sheets, the system comprising:

[0027] The slitting equipment includes an upper cutter and a lower cutter, which are used to slit the initial electrode sheet with double-sided tabs to obtain two electrode sheets with single-sided tabs.

[0028] Four image acquisition units, including two first image acquisition units set on the upper side of the electrode and two second image acquisition units set on the lower side of the electrode, are used to acquire images of the strip edges of the electrode respectively to obtain four initial images;

[0029] The controller is used to acquire the test image corresponding to each initial image; the test image includes the strip edges of the electrode and the electrode region with a preset width based on the strip edges. The test image is obtained by cropping the initial image of the electrode based on the strip edges; the controller detects the strip edges and target edges of the target region in the electrode region, and determines the step detection result based on the distance between the strip edges and the target edges; the step detection result is used to characterize whether the electrode region includes a step; a step is a stepped structure formed on the edge of the electrode due to strip processing; the target edge is any edge on the electrode other than the strip edges; the controller performs aluminum leakage detection on the electrode region in the test image to obtain the aluminum leakage detection result; the aluminum leakage detection result is used to characterize whether the electrode region has aluminum leakage; if the step detection result indicates that the electrode region includes a step, and / or the aluminum leakage detection result indicates that the electrode region has aluminum leakage, the controller determines that the electrode has a defect.

[0030] Thirdly, embodiments of this application provide a defect detection device for electrode sheets, the device comprising:

[0031] The acquisition module is used to acquire the image to be tested. The image to be tested includes the strip edges of the electrode and the electrode area with a preset width based on the strip edges. The image to be tested is obtained by cropping the initial image of the electrode based on the strip edges.

[0032] The first detection module is used to detect the slit edges and target edges of the target area in the electrode region, and to determine the step detection result based on the distance between the slit edges and the target edges. The step detection result is used to characterize whether the electrode region includes steps. A step is a stepped structure formed on the edge of the electrode due to slit processing. The target edge is any edge on the electrode other than the slit edges.

[0033] The second detection module is used to detect aluminum leakage in the electrode area of ​​the image under test and obtain the aluminum leakage detection result; the aluminum leakage detection result is used to characterize whether the electrode area has aluminum leakage.

[0034] The control module is used to determine that there is a defect in the electrode when the step detection result indicates that the electrode area includes a step, and / or the aluminum leakage detection result indicates that there is aluminum leakage in the electrode area.

[0035] Fourthly, embodiments of this application provide a defect detection device for an electrode sheet, including a memory and a processor. The memory stores a computer program that can run on the processor, and the processor executes the program to implement some or all of the steps in the above method.

[0036] Fifthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements some or all of the steps in the above-described method.

[0037] Sixthly, embodiments of this application provide a computer program product, wherein when a computer program or instruction is executed by a processor, it implements some or all of the steps in the above method.

[0038] It should be understood that the above general description and the following detailed description are merely exemplary and explanatory, and are not intended to limit the technical solutions of this application. Attached Figure Description

[0039] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with this application and, together with the specification, serve to explain the technical solutions of this application.

[0040] Figure 1 A schematic diagram of the implementation process of a defect detection method for an electrode sheet provided in this application embodiment. Figure 1 ;

[0041] Figure 2 A schematic diagram of the implementation process of a defect detection method for an electrode sheet provided in this application embodiment. Figure 2 ;

[0042] Figure 3 A schematic diagram of the implementation process of a defect detection method for an electrode sheet provided in this application embodiment. Figure 3 ;

[0043] Figure 4 A top view schematic diagram of an electrode sheet provided in an embodiment of this application;

[0044] Figure 5 A cross-sectional schematic diagram of an electrode sheet provided in an embodiment of this application;

[0045] Figure 6 Provided for the embodiments of this application Figure 5 A magnified view of a portion of region A in the middle;

[0046] Figure 7 A front view schematic diagram of the image acquisition station provided in the embodiments of this application;

[0047] Figure 8 A side view of the image acquisition station provided in an embodiment of this application;

[0048] Figure 9 A schematic diagram of a defect-free electrode sheet provided in an embodiment of this application;

[0049] Figure 10 A schematic diagram of a defective electrode sheet provided in an embodiment of this application;

[0050] Figure 11 A schematic diagram of the hardware entity of a defect detection system for an electrode sheet provided in an embodiment of this application;

[0051] Figure 12 A schematic diagram of the hardware entity of a defect detection device for an electrode sheet provided in an embodiment of this application;

[0052] Figure 13 This is a schematic diagram of the hardware entity of a defect detection device for an electrode sheet provided in an embodiment of this application. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application are further described in detail below with reference to the accompanying drawings and embodiments. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0054] In the following description, references to "some embodiments" refer to a subset of all possible embodiments. It is understood that "some embodiments" may be the same or different subsets of all possible embodiments and may be combined with each other without conflict. The terms "first / second / third" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first / second / third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0055] Unless otherwise defined, 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 pertains. The terminology used herein is for descriptive purposes only and is not intended to limit the scope of this application.

[0056] In related technologies, the slitting process of lithium battery electrodes mainly relies on manual microscopic sampling to detect planar steps and aluminum leakage defects at the electrode edges. This method suffers from low efficiency, high false negative rate, and inability to achieve full inspection, making it difficult to meet the quality control requirements of modern continuous production. To address these issues, this application provides an electrode identification method that automatically acquires images and accurately detects defects, achieving online real-time detection and identification of electrode defects. In this application embodiment, the electrode identification method can be executed by a processor of a computer device. The computer device can refer to a server, laptop, tablet, desktop computer, mobile device, or other device with data processing capabilities.

[0057] Figure 1 A schematic diagram of the implementation process of a defect detection method for an electrode sheet provided in this application embodiment. Figure 1 ,like Figure 1 As shown, the method includes the following steps S101 to S104:

[0058] Step S101: Obtain the image to be tested. The image to be tested includes the strip edges of the electrode and the electrode area with a preset width retained based on the strip edges. The image to be tested is obtained by cropping the initial image of the electrode based on the strip edges.

[0059] Here, the slit edge of the electrode refers to the edge portion of the electrode after it has been cut by the blades of the slitting equipment during the lithium battery production process. During cutting, due to the squeezing action, the slit edge may form defects such as steps or exposed aluminum. The preset width refers to the pre-set width of the electrode area retained based on the slit edge. The electrode area refers to the region within the electrode formed by moving a preset width from the slit edge towards the electrode tab.

[0060] In this embodiment, after the initial electrode sheet is slit by the cutter of the slitting device to obtain the electrode sheet, an image acquisition device located on the upper or lower side of the electrode sheet is used to acquire images of the slit edges of the electrode sheet to obtain an initial image of the electrode sheet. Then, the initial image of the electrode sheet is cropped based on the slit edges to obtain the image to be tested.

[0061] It is understandable that because the slitting edge is generated by the slitting equipment when the original electrode sheet is slitting, step defects and aluminum leakage defects are likely to occur near the slitting edge. Therefore, the area in the initial image other than the area near the slitting edge is redundant data, so the initial image needs to be cropped to remove redundant data and reduce the interference of irrelevant areas on defect detection.

[0062] Step S102: Detect the strip edges and target edges of the target area in the electrode region, and determine the step detection result based on the distance between the strip edges and the target edges; the step detection result is used to characterize whether the electrode region includes steps.

[0063] Here, the steps in the electrode region refer to the stepped structure formed on the edge of the electrode due to the slitting process. The target edge is any edge on the electrode other than the slitting edge.

[0064] In this embodiment, the contour line of the electrode region can be detected first. This contour line can be the striped edge and the target edge of the target region. Based on the distance between the striped edge and the target edge, it can be determined whether the electrode region includes a step. It is understood that because the step in the electrode region is a stepped structure, the surface of the step is not on the same plane as the surface of the electrode body region in the electrode region, resulting in redundant contour lines (i.e., target edges) in the electrode region. Therefore, the step detection result can be determined based on the distance between the striped edge and the target edge.

[0065] In some embodiments, a trained step detection model can be used to detect steps in the electrode regions of the image under test to obtain step detection results. The training process of this trained step detection model may include: constructing a sample set; the sample set includes image information of electrodes with different types of step defects, and each image contains a label indicating the step defect; and training the step detection model using this sample set to obtain the trained step detection model. For example, the step detection model may be a neural network model.

[0066] Step S103: Perform aluminum leakage detection on the electrode area in the image to be tested to obtain the aluminum leakage detection result; the aluminum leakage detection result is used to characterize whether the electrode area leaks aluminum.

[0067] The aluminum leakage detection in this embodiment is performed during the die-cutting and slitting process of the electrode sheet. This differs from the aluminum leakage detection process performed on the electrode sheet surface after the aluminum foil coating process is completed in that:

[0068] First, the causes of aluminum leakage defects differ. In the die-cutting and slitting process, aluminum leakage occurs because the cutting tools of the slitting equipment, due to the squeezing action, may cause aluminum leakage defects at the slitting edge. For example... Figure 4 As shown, the inspection target for aluminum leakage detection in the die-cutting and slitting process is region 406; while the reasons for aluminum leakage defects in the electrode sheets formed after the aluminum foil coating process are: the electrode surface may be scratched during the production process, or aluminum leakage defects may occur due to uneven coating and / or incomplete coating during aluminum foil coating, such as... Figure 4 As shown, the detection target for aluminum leakage detection in the die-cutting and slitting process is area 407.

[0069] Secondly, the detection accuracy for aluminum leakage defects differs. Because the area of ​​aluminum leakage at the edge of the electrode slitting process is small, the image acquisition device used to capture the original image requires high resolution. For example, the detection accuracy for aluminum leakage defects in the die-cutting slitting process needs to reach the micrometer level; for instance, it can detect aluminum leakage defects with an area of ​​50 μm. However, the detection accuracy for aluminum leakage defects after the aluminum foil coating process is lower because defects such as scratches, uneven coating, and missed coating on the electrode surface result in a larger area of ​​aluminum leakage. Therefore, the resolution requirement for the image acquisition device used to capture the image is lower. For example, the detection accuracy for aluminum leakage defects in the die-cutting slitting process is at the millimeter level; for instance, it can detect aluminum leakage defects with an area of ​​3 mm.

[0070] In this embodiment, since the pixel values ​​of the coating on the aluminum foil and the electrode sheet are different in the image, the normal electrode coating area and the aluminum foil area can be distinguished in the electrode area of ​​the image to be tested by using a preset pixel value threshold, thereby completing the detection of aluminum leakage in the electrode area.

[0071] For example, aluminum foil is a metal that exhibits a metallic luster in the image to be tested, while the coating on the electrode surface is slightly darker in color. Therefore, the difference in pixel values ​​between the two can accurately detect whether there is a defect of aluminum leakage in the electrode area.

[0072] Step S104: If the step detection result indicates that the electrode area includes a step, and / or the aluminum leakage detection result indicates that the electrode area has aluminum leakage, then it is determined that there is a defect in the electrode.

[0073] In this embodiment of the application, when the step detection result indicates that the electrode area includes a step, and / or the aluminum leakage detection result indicates that the electrode area has aluminum leakage, a defect detection result can be determined that the electrode has a defect. The defect can be a step defect and / or an aluminum leakage defect.

[0074] In some embodiments, when a defect exists in the electrode sheet, the electrode sheet defect detection device can control a marking device to mark the electrode sheet corresponding to the initial image. In this embodiment, the electrode sheet defect detection device sends a first control signal to the marking device, and the marking device responds to the first control signal to mark the electrode sheet corresponding to the initial image. The first control signal can be a no-good (NG) signal for the electrode sheet; for example, the first control signal can be a low-level signal.

[0075] In some embodiments, if the step detection result indicates that the electrode area does not include a step, and the aluminum leakage detection result indicates that the electrode area does not have aluminum leakage, a defect detection result indicating that the electrode has no defects can be determined. In this case, the electrode defect detection device can send a second control signal to the marking device or not send a signal to the marking device. The marking device responds to the second control signal by not marking the electrode. The second control signal can be a pass / fail signal for the electrode; for example, the second control signal can be a high-level signal.

[0076] In this embodiment, the image to be tested, firstly, is obtained by cropping the edges to remove redundant data, thereby reducing interference from irrelevant areas in defect detection and improving detection efficiency. Secondly, by combining step detection and aluminum leakage detection methods, the presence of steps or aluminum leakage is determined, thus achieving multi-dimensional detection of electrode quality. Compared to the manual sampling method in related technologies, the electrode identification method provided in this embodiment not only improves detection accuracy but also increases detection efficiency.

[0077] In some embodiments, such as Figure 2 As shown, step S103 can be achieved through steps S201 to S203:

[0078] Step S201: The trained aluminum leakage detection model is used to perform the first aluminum leakage detection on the electrode area in the image to be tested, and the first detection result is obtained.

[0079] In this embodiment, the aluminum leakage detection model is a detection network built based on a deep learning algorithm, used to identify whether aluminum leakage defects exist at the edges of electrode strips. This trained aluminum leakage detection model is trained using a large number of labeled electrode images, enabling it to automatically extract image features and output bounding boxes and category prediction information. The process of using the aluminum leakage detection model to make a preliminary judgment on the electrode region in the image is called the first aluminum leakage detection, which is used to identify whether aluminum leakage occurs.

[0080] For example, the aluminum leakage detection model can be a YOLOv5 model, which includes a Backbone network, a Neck network, and a Head network. The Backbone network extracts image features from the image to be tested, the Neck network fuses these image features, and the Head network predicts the fused image features, outputting a first detection result.

[0081] Step S202: Reconstruct the image to be tested and determine the reconstruction error between the reconstructed image and the image to be tested.

[0082] Understandably, when the cutting blades of a slitting machine cut the electrode sheets, the shapes of aluminum leakage defects appearing on the slitting edges due to the extrusion effect can vary greatly. Therefore, when training the aluminum leakage detection model, it is impossible to construct sample images that can include all types of aluminum leakage defects. Consequently, the first aluminum leakage detection may miss some defects. Therefore, to further improve the accuracy of aluminum leakage detection, a second aluminum leakage detection is needed on the image under test. This involves reconstructing the image under test to detect whether aluminum leakage exists.

[0083] In this embodiment, a neural network model can be used to reconstruct the image under test. This neural network model can be obtained through unsupervised training using sample images of electrode sheets without defects. After reconstructing the image under test, the reconstruction error between the reconstructed image and the image under test can be determined. This reconstruction error can determine whether there is an aluminum leakage defect in the electrode sheet region of the image under test. It is understood that because normal areas of the electrode sheet (such as uniform coatings, smooth edges, and surfaces free of foreign matter) have stable texture and grayscale distribution, the neural network model can effectively learn and accurately reconstruct them; however, aluminum leakage defects disrupt this regularity, causing differences between the reconstructed image and the image under test. Therefore, the reconstruction error can be used to determine whether there is an aluminum leakage defect in the electrode sheet region of the image under test.

[0084] Step S203: If the first detection result indicates that aluminum leakage exists in the electrode area, and / or the reconstruction error does not meet the conditions, the aluminum leakage detection result is determined to indicate that aluminum leakage exists in the electrode area.

[0085] In this embodiment, the trained aluminum leakage detection model can output a label indicating whether aluminum leakage exists in the electrode area. The reconstruction error not meeting the condition can mean that the reconstruction error is greater than or equal to a preset value. That is, when the label output by the trained aluminum leakage detection model indicates that aluminum leakage exists in the electrode area, and / or the reconstruction error is greater than or equal to a preset value, the aluminum leakage detection result is determined to be that aluminum leakage exists in the electrode area.

[0086] For example, such as Figure 6 As shown, region 602 is the aluminum leakage region in the electrode region.

[0087] In this embodiment, a trained aluminum leakage detection model is used to detect aluminum leakage in the image under test, improving the detection accuracy. Simultaneously, reconstruction processing is combined with aluminum leakage detection on the image under test, reducing the possibility of missed detections. Compared to a single detection method, this embodiment improves the robustness of aluminum leakage detection through multi-criteria fusion.

[0088] In some embodiments, step S202 can be implemented by steps S2021 to S2023:

[0089] Step S2021: Perform dimensional compression processing on the image to be tested to obtain a multidimensional feature vector.

[0090] Here, dimensionality compression is a processing method that extracts key features from an image using dimensionality reduction algorithms. For example, it can be implemented using models such as Principal Component Analysis (PCA), autoencoders, or Convolutional Neural Networks (CNNs).

[0091] In this embodiment, dimensionality compression removes redundant information from the image while retaining key data that characterizes the essential features of the image, thereby reducing subsequent computational complexity and improving detection efficiency. The multidimensional feature vector obtained after dimensionality compression of the image contains the main structural and texture information of the image, significantly reducing the data size without losing key details. Dimensionality compression also helps improve the speed and accuracy of subsequent reconstruction processing.

[0092] In some embodiments, the image to be tested can be scaled first, and then the scaled image can be cropped to obtain multiple sub-images to be tested. Finally, for each sub-image to be tested, pixel normalization is performed to obtain the input image with dimensionality compression.

[0093] For example, the size of the image to be tested is 4096×290, that is, the shorter side of the image to be tested is 290. The scaling ratio of the image to be tested is determined as: shorter side size / scaling size = 290 / 256 ≈ 0.8828. The 4096×290 image is scaled by the scaling ratio to obtain a scaled image of 3616×256. Then, the scaled image is cropped to obtain multiple sub-images of 256×256. The pixel value range of the sub-images is mapped from [0, 255] to [0, 1] to complete the pixel normalization process. Finally, the 256×256 sub-images are input into the autoencoder, which compresses the 256×256 sub-images into a 256-dimensional low-dimensional multidimensional feature vector.

[0094] Step S2022: Reconstruct the multidimensional feature vectors to obtain the reconstructed image.

[0095] Here, reconstruction processing refers to using multi-dimensional feature vectors to regenerate a reconstructed image that approximates the image to be tested.

[0096] In this embodiment of the application, a multidimensional feature vector can be input into a decoder, and the decoder reconstructs the multidimensional feature vector into an image, that is, reconstructs the image.

[0097] Step S2023: The mean square error of the reconstructed image relative to the image to be tested is determined as the reconstruction error.

[0098] Here, Mean Squared Error (MSE) is used to measure the pixel-level difference between the reconstructed image and the image under test. The smaller the MSE value, the closer the reconstructed image is to the image under test, and the higher the image quality; conversely, the larger the MSE value, the greater the difference between the reconstructed image and the image under test, indicating that there may be anomalies or defects in the image under test.

[0099] Understandably, reconstruction processing can restore the basic shape and main features of the image under test, and the decoder can distinguish between normal and abnormal image features under unsupervised conditions. For example, if the image under test has defects, the reconstructed image will show obvious differences because it cannot accurately restore the defective area. The pixel-level difference between the reconstructed image and the original image can be evaluated by quantitative metrics (such as mean square error).

[0100] In this embodiment of the application, the mean square error of the reconstructed image relative to the image to be tested can be achieved by formula (1):

[0101] (1);

[0102] Where N is the total number of pixels in the image. and These are the i-th pixel values ​​of the original image and the reconstructed image, respectively.

[0103] In this embodiment, a method for extracting features through dimensionality reduction and reconstructing the image is used to extract the core features of the image. Mean squared error is then used to measure the image reconstruction quality, thereby effectively distinguishing between normal and abnormal images. This can compensate for missed detections caused by using a trained aluminum leakage detection model, thus improving the accuracy of aluminum leakage detection.

[0104] In some embodiments, such as Figure 3 As shown, step S102 can be achieved through steps S301 to S303:

[0105] Step S301: Based on a preset grayscale threshold, separate the electrode body region from the target region in the electrode region to obtain the target region.

[0106] Here, the preset grayscale threshold refers to a range of grayscale values ​​pre-defined during image processing to distinguish different regions in the image under test. For example, in the image under test, the electrode body region (i.e., the coating layer) typically has a lower grayscale value, while the target region exhibits a higher grayscale value. By setting an appropriate grayscale threshold, the image can be divided into two parts: one part is the electrode body region, and the other part is the target region, thus providing an accurate image basis for subsequent edge detection. The target region can be the area in the electrode region where the probability of step defects is highest.

[0107] It is understandable that when there are steps in the electrode area, since the surface of the steps and the surface of the electrode body area are not on the same plane, the two will appear as different colors in the image of the electrode to be tested. Therefore, the area with steps can be distinguished from the electrode body area by using a preset grayscale threshold.

[0108] In this embodiment, using a preset grayscale threshold for region division can effectively reduce the influence of background noise, thereby more accurately identifying potential defective regions.

[0109] Step S302: Detect the strip edges and target edges in the target area to obtain the position information of multiple first pixels of the strip edges and the position information of multiple second pixels of the target edges; the pixel width of the strip edges and target edges in the step area is one pixel.

[0110] Here, the slit edge refers to the original cut edge formed after the original electrode sheet has undergone slitting processing, that is, the edge line after the original electrode sheet is cut by the slitting equipment. The target edge is the edge formed near the slit edge due to the squeezing of the cutting tool of the slitting equipment. This edge may be an abnormal structure caused by reasons such as cutter wear.

[0111] In this embodiment, the position information of the first pixel of the strip edge and the position information of the second pixel of the target edge can represent the position information of the first pixel and the second pixel in the image to be measured. The pixel width of the strip edge and the target edge is one pixel, which means that the detected edge line occupies only one pixel width. This can help improve the accuracy of subsequent measurements and reduce errors caused by edge blurring or multi-pixel width.

[0112] For example, such as Figure 4 and Figure 5 As shown, the electrode body region 401 and the target region 402 are separated by a preset grayscale threshold. Then, the strip edge 403 and the target edge 404 in the target region 402 are detected to obtain the position information of multiple first pixels of the strip edge 403 and the position information of multiple second pixels of the target edge 404.

[0113] In practical applications, the detection of strip edges and target edges is usually achieved through edge detection operators (such as Canny, Sobel, etc.).

[0114] Step S303: Based on the position information of multiple first pixels and multiple second pixels, determine the distance between the strip edge and the target edge, and determine the step detection result based on the distance between the strip edge and the target edge.

[0115] In this embodiment, the presence of a step defect can be determined by the width information between the strip edge and the target edge (i.e., the distance between the strip edge and the target edge). Therefore, the width information between the strip edge and the target edge can be determined based on the position information of multiple first pixels and multiple second pixels.

[0116] In some embodiments, the average position information of multiple first pixels can be determined based on the position information of multiple first pixels, and the average position information of multiple second pixels can be determined based on the position information of multiple second pixels. Then, the distance information between the average position information of multiple first pixels and the average position information of multiple second pixels can be determined as the width information between the strip edge and the target edge.

[0117] In this embodiment of the application, if the width information between the strip edge and the target edge is greater than or equal to the preset width, the step detection result is determined to be that the electrode area includes a step.

[0118] In this embodiment, the separation of the electrode body area and the target area is achieved by setting a grayscale threshold, which helps to accurately locate the target area where there may be steps; then, the precise position information of the strip edge and the target edge is obtained by pixel-level edge detection, thereby providing a reliable basis for determining the width information of the target area.

[0119] In some embodiments, step S303 can be implemented by steps S3031 to S3034:

[0120] Step S3031: Based on the position information of multiple first pixels and multiple second pixels, determine multiple pixel pairs including a first pixel and a second pixel; the first pixel and the second pixel in each pixel pair are in the same pixel column or the same pixel row.

[0121] In this embodiment, a first pixel and a second pixel located in the same pixel column or the same pixel row can be defined as a pixel pair. Each pixel pair represents pixels on either side of a step edge that may exist in the electrode region.

[0122] For example, if multiple first pixels include pixel 1 (x1, y1), pixel 2 (x1, y2), and pixel 3 (x1, y3), and multiple second pixels include pixel 4 (x2, y1), pixel 5 (x2, y2), and pixel 6 (x2, y3), then (pixel 1, pixel 4), (pixel 2, pixel 5), and (pixel 3, pixel 6) are considered as three pixel pairs. In this case, the first and second pixels in each pixel pair are in the same pixel row. If multiple first pixels include pixel 1 (x1, y1), pixel 2 (x2, y1), and pixel 3 (x3, y1), and multiple second pixels include pixel 4 (x1, y2), pixel 5 (x2, y2), and pixel 6 (x3, y2), then (pixel 1, pixel 4), (pixel 2, pixel 5), and (pixel 3, pixel 6) are considered as three pixel pairs. In this case, the first and second pixels in each pixel pair are in the same pixel column.

[0123] Step S3032: For each pixel pair, determine the pixel distance between the first pixel and the second pixel based on the position information of the first pixel and the position information of the second pixel in the pixel pair.

[0124] Here, pixel distance refers to the distance between two pixels in an image. For example, the pixel distance can be at least one of the spatial distance between the first pixel and the second pixel, the Euclidean distance, and the Manhattan distance.

[0125] Step S3033: The maximum pixel distance among the pixel distances corresponding to multiple pixel pairs is determined as the width information of the target region in the polarimetric region.

[0126] The width information is used to represent the distance between the strip edge and the target edge.

[0127] In this embodiment of the application, the distance value of the pixel pair with the largest distance between two pixels can be determined as the width information of the target region in the polarimetric region.

[0128] For example, such as Figure 6 As shown, the width information 601 of the target area is the distance information between the strip edge 403 and the target edge 404.

[0129] Step S3034: If the width information of the target area is greater than or equal to the preset width, determine that the step detection result is that the electrode area includes the step.

[0130] In this embodiment, the preset width is a critical value set according to process standards and testing requirements, used to determine whether a step defect exists in the image to be tested. When the width of the target area is greater than or equal to the preset width, it is determined that a step defect exists in the target area, and the corresponding detection result is output.

[0131] For example, such as Figure 6 As shown, when the width information 601 is greater than or equal to the preset width, the width information 601 is determined to be the width information of the step, that is, the electrode area includes the step.

[0132] In this embodiment, the width of the step can be quantified through pixel pair matching and pixel distance calculation, thereby achieving automatic measurement of the step size. Compared with traditional manual visual inspection, this method not only improves the detection accuracy but also provides real-time feedback of the detection results, offering data support for process adjustments.

[0133] In some embodiments, the above method can also be implemented through steps S11 to S13:

[0134] Step S11: Obtain the initial image of the electrode sheet; the electrode sheet is a single-sided electrode sheet obtained by slicing the initial electrode sheet with both sides of the electrode tab.

[0135] Here, "double-sided tabs" refers to electrode sheets with conductive metal components (such as aluminum foil) on both sides for electrical connection. Slitting is a process of cutting electrode sheets with double-sided tabs into single-sided tabs, typically done using slitting equipment. Slitting divides a complete electrode sheet into multiple independent single electrode sheets to meet subsequent assembly requirements. After slitting, the original electrode sheet with double-sided tabs retains only one side of the tab, while the other side is cut off, thus forming the processed single-sided tab electrode sheet.

[0136] In this embodiment of the application, during the die-cutting and slitting process, the slitting equipment first slits the initial electrode sheet, which includes double-sided tabs, using a cutting tool to obtain an electrode sheet including a single-sided tab. Specifically, one initial electrode sheet with double-sided tabs can be slit into two electrode sheets with single-sided tabs. Then, image acquisition units located above and below the electrode sheet acquire images of the slitting edges of the electrode sheet to obtain initial images of the electrode sheet.

[0137] For example, such as Figure 4 As shown, the electrode 400 obtained after slitting includes two single-sided tabs 405.

[0138] Step S12: Perform edge detection on the initial image.

[0139] In this embodiment, strip edge detection refers to using image processing algorithms to identify edge regions generated by striping in the initial image. Since striping leaves obvious edge traces on the image, these traces manifest as specific grayscale variations or texture features. By applying edge detection operators (such as Canny, Sobel, etc.) or deep learning models, the initial image can be processed and the positions of the strip edges can be located. Accurate strip edge detection has a significant impact on subsequent image cropping processes, as it determines the boundary range of the area to be measured and affects the detection accuracy.

[0140] For example, such as Figure 4 As shown, by performing strip edge detection on the initial image, strip edge 403 in the initial image is detected.

[0141] Step S13: Based on the strip edges and the cropping edges whose distance to the strip edges is within a preset distance, the detected initial image is cropped to obtain the image to be tested.

[0142] In this embodiment of the application, after detecting the strip edge, a cropping edge can be obtained by extending a preset distance inward from the strip edge along the tab direction. Then, based on the strip edge and the cropping edge, the detected initial image is cropped to obtain the image to be tested.

[0143] For example, if the initial image size is 4096*3000, then by cropping the initial image, a test image with a size of 4096*290 can be obtained.

[0144] In this embodiment, an initial image of the initial electrode sheet of both sides of the electrode is obtained after striping. Then, a striping edge detection operation is performed, and a cutting edge is determined based on a preset distance. The initial image is then cropped to generate a more accurate image to be tested. This method effectively reduces the detection range, improves the accuracy of defect identification, and thus enhances overall detection efficiency.

[0145] In some embodiments, the above method is applied to the die-cutting and slitting process; the initial electrode sheet after slitting includes an electrode sheet with two single-sided tabs; the above method can also be implemented through steps S21 and S22:

[0146] Step S21: For each electrode sheet of a single-sided electrode, acquire two initial images obtained by acquiring the strip edges of the electrode sheet using two image acquisition devices.

[0147] In this embodiment, each electrode sheet on one side of the electrode corresponds to two image acquisition devices, which are respectively distributed on the upper and lower sides of the electrode sheet. After the initial electrode sheet is stripped, images are acquired from the stripped edges of the initial electrode sheet. For example, the image acquisition device can be a camera.

[0148] In practical applications, on power lithium battery production lines, when the electrode sheet of a single-sided tab reaches the shooting position of the image acquisition unit, the system will trigger two image acquisition units to simultaneously capture images in real time. These two image acquisition units are installed on the upper and lower sides of the electrode sheet's running path, respectively. One image acquisition unit faces side A of the electrode sheet, and the other faces side B of the electrode sheet. These two image acquisition units work together to ensure the synchronicity and consistency of image acquisition.

[0149] Figure 7 This is a front view schematic diagram of the image acquisition station provided in an embodiment of this application. Figure 8 This is a side view of the image acquisition station provided in an embodiment of this application, as shown below. Figure 7 and Figure 8 As shown, the electrode 400 has a first image acquisition unit 701 and a second image acquisition unit 702. The first image acquisition unit 701 is located on the upper side of the electrode 400, and the second image acquisition unit 702 is located on the lower side of the electrode 400. When the electrode 400 is located at the acquisition position of the corresponding image acquisition unit, the first image acquisition unit 701 and the second image acquisition unit 702 acquire the initial image of the electrode. Specifically, the first image acquisition unit 701 is the image acquisition unit facing side A of the electrode, and the second image acquisition unit 702 is the image acquisition unit facing side B of the electrode.

[0150] Step S22: Perform edge detection and cropping on each initial image to obtain the image to be tested.

[0151] In this embodiment of the application, each image acquisition device acquires an initial image of an electrode. Each initial image can be segmented and edge detected. Then, based on the segmented edges and the cropping edges whose distance to the segmented edges is within a preset distance, the detected initial image is cropped to obtain the image to be tested.

[0152] In this embodiment, by employing two image acquisition devices to initially acquire images of the upper and lower sides of the electrode sheet on one side of the electrode tab during the die-cutting and slitting process, and combining this with the operation of slitting edge detection and trimming, efficient and accurate detection of slitting edge defects in the electrode sheet on one side of the electrode tab is achieved. This operation method ensures that both the upper and lower sides of the electrode sheet on one side of the electrode tab can be fully detected, thereby avoiding missed detections and improving the comprehensiveness of the electrode sheet inspection.

[0153] In some embodiments, the above method can also be implemented through steps S31 and S32:

[0154] Step S31: If the step detection result indicates that the electrode area includes a step, and / or the aluminum leakage detection result indicates that the electrode area has aluminum leakage, then the defect detection result of the electrode is determined to be unqualified.

[0155] Step S32: If the defect detection results of multiple consecutive electrode sheets are unqualified, a prompt message indicating that the cutting tool of the slitting equipment is abnormal is sent to the slitting equipment.

[0156] In this embodiment, the slitting equipment refers to the equipment responsible for cutting initial electrode sheets with double tabs into single-tab electrode sheets, and its core component is the cutting tool. Steps or aluminum leakage defects occur because the cutting tool needs to compress the initial electrode sheet during the cutting process to complete the slitting; when the cutting tool malfunctions, the likelihood of defects appearing on the slitting edge increases. Therefore, when the defect detection results of multiple consecutive electrode sheets are all unqualified, it indicates that the cutting tool of the slitting equipment is abnormal, and a tool abnormality warning message needs to be sent to the slitting equipment. For example, the number of electrode sheets identified as unqualified can be 3. That is, when the defect detection results of 3 consecutive electrode sheets are unqualified, a warning message indicating a tool abnormality in the slitting equipment is sent to the slitting equipment.

[0157] In this embodiment of the application, by setting up continuous NG judgment logic, the equipment status can be fed back in a timely manner when continuous defects occur, which helps operators to quickly find the cause of the fault, prevent the generation of batch defective products, and improve production stability.

[0158] The following describes the application of the electrode defect detection method provided in the embodiments of this application in a real-world scenario.

[0159] In the production of power lithium batteries, electrode slitting is one of the key processes. Due to process and equipment limitations, the squeezing action of the cutting tools on the electrodes during slitting can cause planar steps to form at the cut edges, which can severely expose the current collector (aluminum foil). Planar step aluminum leakage poses significant safety hazards: firstly, it makes the stepped area prone to puncturing the separator, leading to a short circuit between the cathode and anode; secondly, the exposed current collector can cause internal short circuits, ultimately resulting in serious safety issues such as cell fires. Therefore, effectively detecting and reducing planar step aluminum leakage defects is a core issue for improving battery safety and product quality.

[0160] For example, such as Figure 9 As shown, no step defects or aluminum leakage defects were observed on the slit edge 403 of electrode 400. Figure 10 As shown, the slit edge 403 of the electrode 400 exhibits step defects 1001 and aluminum leakage defects 1002.

[0161] In the production of power lithium batteries, the traditional method for detecting planar step aluminum leakage defects in the electrode slitting process relies on manual sampling. Operators need to periodically cut electrode samples of a certain length from production equipment such as die-cutting machines, slitting machines, and stacking machines. These samples are then placed under a metallographic microscope or electron microscope for microscopic observation under high magnification. Through microscopic observation, technicians can identify the specific type of defect, such as damaged cutters or improper cutter gaps. They can also use the measuring tools provided with the microscope to accurately measure the dimensional parameters of the defect, including the length and width of the aluminum leakage area.

[0162] However, manual inspection methods have the following significant limitations: low inspection efficiency, unable to achieve full inspection of electrode sheets; inspection results rely on the experience and subjective judgment of technicians, resulting in significant individual differences; furthermore, manual operation cannot meet the needs of modern continuous production, making it difficult to achieve real-time and efficient defect detection. As lithium battery production develops towards intelligence, automation, and high precision, traditional manual inspection methods can no longer meet the requirements of high-quality and high-efficiency production. There is an urgent need for an online inspection technology based on machine vision and artificial intelligence to address the problem that traditional manual inspection methods cannot meet the requirements of high-quality and high-efficiency production.

[0163] To address the technical problems in the aforementioned related technologies, embodiments of this application provide a method for detecting defects in electrode sheets. This identification method can be implemented through steps S41 to S44:

[0164] In step S41, the electrode defect detection device performs image preprocessing on the electrode image and performs step detection on the processed electrode image to obtain the first step detection result.

[0165] In this embodiment of the application, the vision host computer acquires the electrode image captured by the image acquisition device and transmits it to the electrode defect detection device equipped with the detection algorithm. The electrode defect detection device processes the electrode image.

[0166] In this embodiment of the application, step S41 can be implemented by steps S411 to S445:

[0167] Step S411: Crop the polarimetric image to obtain the first image.

[0168] In this embodiment, the polarimetric image has a large resolution (e.g., 4096*3000) and a small effective detection area. The cropped first image (e.g., 4096*290) is used to reduce the image size and the amount of data processed.

[0169] In this embodiment, the edge detection algorithm Canny is used to coarsely locate the contour edges in the polarimetric image, determine the coordinate values ​​to be cropped, and the Region of Interest (ROI) algorithm is used to crop the image to obtain the first image.

[0170] Step S412: Denoise the first image to obtain the second image.

[0171] In this embodiment, a bilateral filtering algorithm is used to denoise the first image, which can more accurately preserve edge details while smoothing noise (through dual weight control of spatial domain and grayscale domain), and avoid noise being misjudged as edges.

[0172] Step S413: Perform histogram equalization on the second image to obtain the third image.

[0173] In this embodiment, histogram equalization is used to improve the grayscale difference between the target and the background in the second image.

[0174] Step S414: Separate the target in the third image from the background using a global threshold to obtain a binarized image.

[0175] Step S415: Extract the contour edge of the target in the binarized image to obtain a single-pixel wide edge line.

[0176] Step S416: Calculate the maximum distance between the point sets corresponding to the two sets of edge lines, and determine the step detection result based on the maximum width.

[0177] In this embodiment of the application, when the maximum width is greater than the preset width, the step detection result of the electrode sheet is determined to have a step.

[0178] In step S42, the electrode defect detection device uses the YOLOv5 model to detect the area of ​​aluminum leakage in the electrode image and obtain the aluminum leakage detection result.

[0179] In this embodiment, the polarimetric image is first scaled proportionally along its shorter side, and then zeros are added to any unfilled portions to determine the final input image size for the YOLOv5 model (320*320). Simultaneously, the image is normalized, mapping the pixel value range from [0, 255] to [0, 1] to prevent large numerical ranges from affecting model training. The adjusted image is then converted into Tensor data and input into the YOLOv5 model.

[0180] In this embodiment, the YOLOv5 target detection network identifies the electrode detection region (the YOLOv5 network structure is as follows). Figure 2 (As shown). YOLO v5 mainly consists of four parts: the input layer, the backbone network, the Neck, and the Prediction layer. Each part is explained below:

[0181] Input end: Performs image enhancement operations such as cropping, splicing, and affine transformation on the input image.

[0182] Backbone: A convolutional neural network that aggregates and forms image features at different fine-grained levels of images.

[0183] Neck: A series of network layers that mix and combine image features and pass the image features to the prediction layer.

[0184] Head (Prediction): Predicts image features, generates bounding boxes, and predicts categories.

[0185] In this embodiment, the YOLOv5 model first extracts image features through a Backbone network. The feature maps are then fused using a Neck network to enhance detection performance. The decoder converts the feature maps into predicted bounding boxes and class probabilities.

[0186] In this embodiment, the YOLOv5 model filters out detection results with low confidence based on a preset confidence threshold of 0.8 (continuously monitored and updated), and outputs confidence results that meet the threshold requirement. The model also outputs the category label of aluminum leakage defects, the bounding box coordinates of the defects, and visualizes them on the original image for easy and intuitive viewing of the results.

[0187] In step S43, the electrode defect detection device reconstructs the electrode image and performs step detection on the electrode image based on the reconstructed image to obtain the second step detection result.

[0188] In this embodiment, the autoencoder compresses the input image into a 256-dimensional low-dimensional feature vector, and the decoder reconstructs the feature vector into an image. The mean square error is used to calculate the average of the squared differences between the input image and the reconstructed image. Defective images (aluminum leakage, unique morphology) usually lead to large reconstruction errors. A threshold of 0.035 is set (continuously monitored and updated), and the difference between the input image and the reconstructed image is calculated. If the error exceeds the threshold, it is judged as abnormal.

[0189] In step S44, if at least one of the results from the first step detection, the second step detection, and the aluminum leakage detection indicates that the electrode sheet has a defect, the vision computer determines that the electrode sheet is unqualified.

[0190] In this embodiment, the vision-based host computer saves the detection result data (time, step size, aluminum leakage threshold, etc.) and images to the local PC disk for traceability. The host computer generates control signals and sends them to the main equipment of the die-cutting and slitting machine. If any NG products are generated, the equipment alarms but does not stop. The marking machine affixes a label to the defect location. If 3+ NG products appear consecutively, the equipment alarms and stops. The operator checks the equipment, and the tagged NG products are discharged as waste on the winding machine.

[0191] In this embodiment, a complete industrial vision inspection system is constructed to achieve efficient and accurate detection of defects on the electrode plane. This method uses an industrial camera to acquire electrode images and combines them with deep learning algorithms, enabling the entire process from image acquisition to defect detection to be completed within 10ms, ensuring that the detection speed meets the real-time requirements of industrial production. The inspection system can not only accurately identify defects such as aluminum leakage and steps on the electrode plane, but also calculate the specific length parameters of the defects, achieving quantitative analysis of the defects. Finally, the detection results are fed back to the charge-coupled device (CCD) control system in real time in NG / OK format, providing a reliable basis for defect handling during the production process.

[0192] In this embodiment, an industrial camera is used to capture real-time images of the planar portion of the electrode during its slitting process, obtaining electrode images. Because a double-sided slitting blade is used during electrode slitting, the pressure from the blades may cause stepped aluminum leakage defects on both sides of the electrode, making it impossible for a single camera to fully capture defect information from both sides. To address this issue, this invention innovatively employs a dual-camera system to capture images of both sides of the electrode from different angles, ensuring complete acquisition of electrode defect information.

[0193] Figure 11 A defect detection system for an electrode sheet is provided in the embodiments of this application, such as Figure 11 As shown, the defect detection system 1100 for this electrode includes:

[0194] Slitting equipment 1101 includes an upper cutter and a lower cutter, which are used to slit the initial electrode sheet with double-sided tabs to obtain an electrode sheet with two single-sided tabs.

[0195] Four image acquisition units 1102 include two first image acquisition units disposed on the upper side of the electrode sheet and two second image acquisition units disposed on the lower side of the electrode sheet. The two first image acquisition units and the two second image acquisition units are used to acquire images of the strip edges of the electrode sheet respectively to obtain four initial images.

[0196] The controller 1103 is used to acquire a test image corresponding to each initial image; the test image includes the strip edges of the electrode and an electrode region with a preset width based on the strip edges, and the test image is obtained by cropping the initial image of the electrode based on the strip edges; step detection is performed on the electrode region in the test image to obtain a step detection result; the step detection result is used to characterize whether the electrode region includes a step; aluminum leakage detection is performed on the electrode region in the test image to obtain an aluminum leakage detection result; the aluminum leakage detection result is used to characterize whether the electrode region has aluminum leakage; if the step detection result characterizes that the electrode region includes a step, and / or the aluminum leakage detection result characterizes that the electrode region has aluminum leakage, it is determined that there is a defect in the electrode.

[0197] In this embodiment of the application, the upper cutter and the lower cutter of the slitting device are respectively distributed on the upper and lower sides of the initial electrode sheet, and the upper cutter and the lower cutter have different structures. The upper cutter is used to cut the initial electrode sheet, while the lower cutter is used to support the initial electrode sheet.

[0198] In this embodiment of the application, after slicing an initial electrode sheet to obtain two single-sided electrode sheets, the first image acquisition unit corresponding to each single-sided electrode sheet acquires an image of the upper surface of the electrode sheet, and the second image acquisition unit corresponding to each single-sided electrode sheet acquires an image of the lower surface of the electrode sheet. Therefore, one single-sided electrode sheet corresponds to two initial images.

[0199] Figure 12 A defect detection device for an electrode sheet is provided in the embodiments of this application, such as Figure 12 As shown, the defect detection device 1200 for the electrode includes:

[0200] The acquisition module 1201 is used to acquire the image to be tested, which includes the strip edges of the electrode and the electrode area with a preset width based on the strip edges. The image to be tested is obtained by cropping the initial image of the electrode based on the strip edges.

[0201] The first detection module 1202 is used to detect the strip edges and target edges of the target area in the electrode region, and to determine the step detection result based on the distance between the strip edges and the target edges; the step detection result is used to characterize whether the electrode region includes a step.

[0202] The second detection module 1203 is used to perform aluminum leakage detection on the electrode area in the image to be tested and obtain the aluminum leakage detection result; the aluminum leakage detection result is used to characterize whether the electrode area has aluminum leakage.

[0203] The control module 1204 is used to determine that there is a defect in the electrode when the step detection result indicates that the electrode area includes a step, and / or the aluminum leakage detection result indicates that there is aluminum leakage in the electrode area.

[0204] In some embodiments, the second detection module 1203 is further configured to perform a first aluminum leakage detection on the electrode region in the image to be tested using the trained aluminum leakage detection model to obtain a first detection result; perform reconstruction processing on the image to be tested and determine the reconstruction error between the reconstructed image and the image to be tested; and determine that the aluminum leakage detection result indicates that aluminum leakage exists in the electrode region when the first detection result indicates that aluminum leakage exists in the electrode region and / or the reconstruction error does not meet the conditions.

[0205] In some embodiments, the second detection module 1203 is further configured to perform dimensional compression processing on the image to be tested to obtain a multidimensional feature vector; perform reconstruction processing on the multidimensional feature vector to obtain a reconstructed image; and determine the mean square error of the reconstructed image relative to the image to be tested as the reconstruction error.

[0206] In some embodiments, the first detection module 1202 is further configured to separate the electrode body region and the target region in the electrode region based on a preset grayscale threshold to obtain a target region; detect the strip edges and target edges in the target region to obtain the position information of multiple first pixels of the strip edges and the position information of multiple second pixels of the target edges; the pixel width of the strip edges and target edges in the step region is a single pixel; the target edge is the other edge on the electrode besides the strip edges; determine the distance between the strip edges and the target edges based on the position information of multiple first pixels and the position information of multiple second pixels, and determine the step detection result based on the distance between the strip edges and the target edges.

[0207] In some embodiments, the first detection module 1202 is further configured to determine multiple pixel pairs, each including a first pixel and a second pixel, based on the position information of multiple first pixels and multiple second pixels; the first pixel and the second pixel in each pixel pair are in the same pixel column or the same pixel row; for each pixel pair, the pixel distance between the first pixel and the second pixel is determined based on the position information of the first pixel and the second pixel in the pixel pair; the maximum pixel distance among the pixel distances corresponding to the multiple pixel pairs is determined as the width information of the target area in the electrode region; if the width information of the target area is greater than or equal to a preset width, the step detection result is determined to be that the electrode region includes a step.

[0208] In some embodiments, the acquisition module 1201 is further configured to acquire an initial image of the electrode sheet; the electrode sheet is a single-sided electrode sheet obtained by slicing an initial electrode sheet with double-sided tabs; the defect detection device for the electrode sheet further includes a third detection module and a cutting module, the third detection module being configured to detect the slicing edges of the initial image; the cutting module being configured to cut the detected initial image based on the slicing edges and the cutting edges whose distance to the slicing edges is within a preset distance, to obtain the image to be tested.

[0209] In some embodiments, the initial electrode sheet after stripping includes two electrode sheets with one-sided tabs; the acquisition module 1201 is further configured to acquire two initial images for each electrode sheet with one-sided tabs by acquiring the strip edges of the electrode sheet using two image acquisition devices respectively; the defect detection device for the electrode sheet further includes a detection and cutting module, which is configured to perform strip edge detection and cutting on each initial image to obtain the image to be tested.

[0210] In some embodiments, the above-mentioned electrode defect detection device further includes a determining module and a sending module. The determining module is used to determine that the defect detection result of the electrode is unqualified when the step detection result indicates that the electrode area includes a step, and / or the aluminum leakage detection result indicates that the electrode area has aluminum leakage. The sending module is used to send a prompt message to the slitting equipment indicating that the cutting tool of the slitting equipment is abnormal when there are multiple consecutive electrode defect detection results that are unqualified.

[0211] Based on the foregoing embodiments, this application provides a defect detection device for electrode sheets. The device includes various units and modules included in each unit, which can be implemented by a processor in a computer device; of course, it can also be implemented by specific logic circuits. In the implementation process, the processor can be a central processing unit (CPU), a microprocessor unit (MPU), a digital signal processor (DSP), or a field programmable gate array (FPGA), etc.

[0212] The descriptions of the apparatus embodiments above are similar to those of the method embodiments above, and have similar beneficial effects. In some embodiments, the functions or modules included in the apparatus provided in this application can be used to perform the methods described in the method embodiments above. For technical details not disclosed in the apparatus embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0213] It should be noted that, in the embodiments of this application, if the above-described data processing method is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, or the part that contributes to the related technology, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, mobile hard drives, read-only memory (ROM), magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any specific hardware, software, or firmware, or any combination of hardware, software, and firmware.

[0214] This application provides a computer device including a memory and a processor. The memory stores a computer program that can run on the processor. When the processor executes the program, it implements some or all of the steps in the above-described method.

[0215] This application provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements some or all of the steps in the above-described method. The computer-readable storage medium can be transient or non-transient.

[0216] This application provides a computer program including computer-readable code. When the computer-readable code is run in a computer device, the processor in the computer device performs some or all of the steps in the above-described method.

[0217] This application provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program. When the computer program is read and executed by a computer, it implements some or all of the steps in the above-described method. This computer program product can be implemented specifically through hardware, software, or a combination thereof. In some embodiments, the computer program product is specifically embodied as a computer storage medium; in other embodiments, the computer program product is specifically embodied as a software product, such as a software development kit (SDK), etc.

[0218] It should be noted that the descriptions of the various embodiments above tend to emphasize the differences between them, while their similarities or commonalities can be referred to interchangeably. The descriptions of the above embodiments of the device, storage medium, computer program, and computer program product are similar to the descriptions of the above method embodiments and have similar beneficial effects. For technical details not disclosed in the embodiments of the device, storage medium, computer program, and computer program product of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0219] Figure 13 This is a schematic diagram of a hardware entity of the electrode defect detection device in the embodiments of this application, such as... Figure 13 As shown, the hardware entity of the defect detection device 1300 for the electrode sheet includes: a processor 1301, a communication interface 1302, and a memory 1303, wherein:

[0220] The processor 1301 typically controls the overall operation of the electrode defect detection device 1300, which may implement the electrode defect detection method provided in the embodiments of this application.

[0221] The communication interface 1302 enables the electrode defect detection device 1300 to communicate with other terminals or servers via a network.

[0222] The memory 1303 is configured to store instructions and applications executable by the processor 1301, and can also cache data to be processed or already processed (e.g., image data, audio data, voice communication data, and video communication data) in the processor 1301 and the electrode defect detection device 1300. It can be implemented using flash memory or random access memory (RAM). Data transfer between the processor 1301, the communication interface 1302, and the memory 1303 can be performed via bus 1304.

[0223] This application provides a computer storage medium storing one or more programs that can be executed by one or more processors to implement the steps of the electrode defect detection method as described in any of the above embodiments.

[0224] It should be noted that the descriptions of the storage medium and device embodiments above are similar to the descriptions of the method embodiments above, and have similar beneficial effects. For technical details not disclosed in the storage medium and device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0225] The aforementioned processor can be at least one of the following: Application Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), Controller, Microcontroller, and Microprocessor. It is understood that other electronic devices can also implement the functions of the aforementioned processor, and this application does not specifically limit the specific implementation.

[0226] The aforementioned computer storage media / memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), magnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc, or compact disc read-only memory (CD-ROM), etc.; or it can be various terminals that include one or any combination of the above-mentioned memories, such as mobile phones, computers, tablet devices, personal digital assistants, etc.

[0227] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above steps / processes do not imply a sequential order of execution; the execution order of each step / process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above embodiments of this application are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0228] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0229] The above are merely embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for detecting defects in an electrode sheet, characterized in that, The defect detection method for the electrode sheet includes: Acquire a test image, which includes the slit edges of the electrode sheet and an electrode sheet region with a preset width retained based on the slit edges. The test image is obtained by cropping an initial image of the electrode sheet based on the slit edges; the slit edges are the edge portions formed after the blades of the slitting device slit the initial electrode sheet. Based on a preset grayscale threshold, the electrode body region and the target region in the electrode region are separated to obtain the target region; The strip edges and target edges in the target region are detected to obtain the position information of multiple first pixels of the strip edges and the position information of multiple second pixels of the target edges; the pixel width of the strip edges and the target edges in the target region is a single pixel; Based on the position information of the plurality of first pixels and the position information of the plurality of second pixels, a plurality of pixel pairs, each including a first pixel and a second pixel, are determined; in each pixel pair, the first pixel and the second pixel are in the same pixel column or the same pixel row. For each pixel pair, the pixel distance between the first pixel and the second pixel is determined based on the position information of the first pixel and the position information of the second pixel in the pixel pair. The maximum pixel distance among the pixel distances corresponding to the multiple pixel pairs is determined as the width information of the target region in the polarimetric region; the width information is used to characterize the distance between the strip edge and the target edge; If the width of the target area is greater than or equal to a preset width, the step detection result is determined to indicate that the electrode area includes a step; the step detection result is used to characterize whether the electrode area includes a step; the step is a stepped structure formed on the edge of the electrode due to stripping processing; the target edge is any edge on the electrode other than the stripping edge. Aluminum leakage detection is performed on the electrode region in the image to be tested to obtain aluminum leakage detection results; the aluminum leakage detection results are used to characterize whether the electrode region is leaking aluminum. If the step detection result indicates that the electrode area includes a step, and / or the aluminum leakage detection result indicates that the electrode area has aluminum leakage, then it is determined that the electrode has a defect.

2. The defect detection method for electrode sheets according to claim 1, characterized in that, The step of performing aluminum leakage detection on the electrode region in the image to be tested, and obtaining the aluminum leakage detection result, includes: The trained aluminum leakage detection model is used to perform a first aluminum leakage detection on the electrode region in the image to be tested, and a first detection result is obtained. The image to be tested is reconstructed, and the reconstruction error between the reconstructed image and the image to be tested is determined. If the first detection result indicates that aluminum leakage exists in the electrode area, and / or the reconstruction error does not meet the conditions, the aluminum leakage detection result is determined to indicate that aluminum leakage exists in the electrode area.

3. The defect detection method for electrode sheets according to claim 2, characterized in that, The process of reconstructing the image to be tested and determining the reconstruction error between the reconstructed image and the image to be tested includes: The image to be tested is subjected to dimensionality compression processing to obtain a multidimensional feature vector; The multidimensional feature vectors are reconstructed to obtain a reconstructed image; The mean square error of the reconstructed image relative to the image under test is defined as the reconstruction error.

4. The defect detection method for electrode sheets according to any one of claims 1 to 3, characterized in that, The defect detection method for the electrode sheet also includes: Obtain the initial image of the electrode sheet; the electrode sheet is a single-sided electrode sheet obtained by slicing the initial electrode sheet with double-sided tabs. Perform edge detection on the initial image; Based on the strip edges and the cropping edges whose distance to the strip edges is within a preset distance, the detected initial image is cropped to obtain the image to be tested.

5. The defect detection method for electrode sheets according to claim 4, characterized in that, It is used in the die-cutting and slitting process; the initial electrode sheet after slitting includes an electrode sheet with two single-sided tabs; The defect detection method for the electrode sheet also includes: For each of the aforementioned single-sided electrode tabs, two initial images are obtained by acquiring the strip edges of the electrode tab using two image acquisition devices respectively; Each initial image is subjected to edge detection and cropping to obtain the image to be tested.

6. The defect detection method for electrode sheets according to any one of claims 1 to 3, characterized in that, The defect detection method for the electrode sheet also includes: If the step detection result indicates that the electrode area includes a step, and / or the aluminum leakage detection result indicates that the electrode area has aluminum leakage, then the defect detection result of the electrode is determined to be unqualified. If multiple consecutive electrode sheets fail the defect test, a notification message indicating a tool malfunction is sent to the slitting equipment.

7. A defect detection system for electrode sheets, characterized in that, The defect detection system for the electrode includes: The slitting equipment includes an upper cutter and a lower cutter, which are used to slit the initial electrode sheet with double-sided tabs to obtain two electrode sheets with single-sided tabs. Four image acquisition devices, including two first image acquisition devices disposed on the upper side of the electrode and two second image acquisition devices disposed on the lower side of the electrode, wherein the two first image acquisition devices and the two second image acquisition devices are used to acquire images of the strip edges of the electrode respectively to obtain four initial images; A controller is configured to acquire a test image corresponding to each initial image; the test image includes the striping edge of the electrode and an electrode region with a preset width retained based on the striping edge, the test image being obtained by cropping the initial image of the electrode based on the striping edge; the striping edge is the edge portion formed after the initial electrode is stripped by the cutter of the striping device; based on a preset grayscale threshold, the electrode body region and the target region in the electrode region are separated to obtain the target region; the striping edge and the target edge in the target region are detected to obtain the position information of multiple first pixels of the striping edge and the position information of multiple second pixels of the target edge; the pixel width of the striping edge and the target edge in the target region is a single pixel; based on the position information of the multiple first pixels and the multiple second pixels, multiple pixel pairs including one first pixel and one second pixel are determined; the first pixel and the second pixel in each pixel pair are in the same pixel column or the same pixel row; for each pixel pair, Based on the position information of the first pixel and the second pixel in the pixel pair, the pixel distance between the first pixel and the second pixel is determined; the maximum pixel distance among the pixel distances corresponding to the multiple pixel pairs is determined as the width information of the target region in the electrode area; the width information is used to characterize the distance between the strip edge and the target edge; if the width information of the target region is greater than or equal to a preset width, the step detection result is determined to indicate that the electrode area includes a step; the step detection result is used to characterize whether the electrode area includes a step; the step is a stepped structure formed on the edge of the electrode due to strip processing; the target edge is any edge on the electrode other than the strip edge; aluminum leakage detection is performed on the electrode area in the image to be tested to obtain an aluminum leakage detection result; the aluminum leakage detection result is used to characterize whether the electrode area leaks aluminum; if the step detection result indicates that the electrode area includes a step, and / or the aluminum leakage detection result indicates that the electrode area has aluminum leakage, the electrode is determined to have a defect.

8. The electrode defect detection system according to claim 7, characterized in that, The controller is also used for: If the electrode has defects, the marking device is controlled to mark the electrode corresponding to the initial image.

9. A defect detection device for electrode sheets, characterized in that, The defect detection device for the electrode sheet includes: The acquisition module is used to acquire the image to be tested, which includes the slit edge of the electrode and an electrode area with a preset width retained based on the slit edge. The image to be tested is obtained by cropping the initial image of the electrode based on the slit edge; the slit edge is the edge portion formed by the blade of the slitting device after slitting the initial electrode. The first detection module is used to separate the electrode body region from the target region in the electrode region based on a preset grayscale threshold to obtain the target region; detect the strip edges and target edges in the target region to obtain the position information of multiple first pixels of the strip edges and the position information of multiple second pixels of the target edges; the pixel width of the strip edges and the target edges in the target region is one pixel; based on the position information of the multiple first pixels and the multiple second pixels, determine multiple pixel pairs including one first pixel and one second pixel; the first pixel and the second pixel in each pixel pair are in the same pixel column or the same pixel row; for each pixel pair, Based on the position information of the first pixel and the second pixel in the pixel pair, the pixel distance between the first pixel and the second pixel is determined; the maximum pixel distance among the pixel distances corresponding to the multiple pixel pairs is determined as the width information of the target region in the electrode area; the width information is used to characterize the distance between the strip edge and the target edge; if the width information of the target region is greater than or equal to a preset width, the step detection result is determined to indicate that the electrode area includes a step; the step detection result is used to characterize whether the electrode area includes a step; the step is a stepped structure formed on the edge of the electrode due to strip processing; the target edge is any edge on the electrode other than the strip edge; The second detection module is used to perform aluminum leakage detection on the electrode area in the image to be tested, and obtain the aluminum leakage detection result; the aluminum leakage detection result is used to characterize whether the electrode area is leaking aluminum. A control module is configured to determine that the electrode has a defect when the step detection result indicates that the electrode area includes a step, and / or the aluminum leakage detection result indicates that the electrode area has aluminum leakage.

10. A defect detection device for an electrode sheet, comprising a memory and a processor, wherein the memory stores a computer program executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the method according to any one of claims 1 to 6.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images