System and method for detecting defects in ceramic coating of battery electrode sheet

Abstract

The application discloses a battery pole piece ceramic coating defect detection system and method, and relates to the technical field of batteries. A specific embodiment of the system comprises: using an encoder to synchronously trigger an optical imaging unit and an eddy current sensor to collect data of a detection area of a battery pole piece to be detected, and to obtain image signals and eddy current signals respectively; a controller analyzes the intensity of the eddy current signals, and if the intensity exceeds that of a signal of an uncoated aluminum foil, it is determined that the coating of the detection area has fallen off; the controller inputs the image signals into an image recognition model to output a preliminary defect judgment result of the detection area of the battery pole piece to be detected; and the controller uses a positive sample label corresponding to the coating falling off of the detection area and an image feature vector corresponding to the preliminary defect judgment result to train the image recognition model, and detects the ceramic coating defects of the battery pole piece by optimizing the trained image recognition model. The embodiment can improve the recognition accuracy of the ceramic coating defects of the battery pole piece.

Description

Technical Field

[0001] This invention relates to the field of battery technology, and in particular to a system and method for detecting defects in the ceramic coating of battery electrodes. Background Technology

[0002] In the manufacturing process of lithium-ion batteries, the white ceramic coating applied to the tab area of ​​the electrode plays a crucial insulating role. If this coating peels off, it will directly expose the underlying aluminum foil, drastically increasing the risk of internal short circuits and posing a serious safety hazard. Figure 1 In the diagram, A is the tab of the electrode, B is the ceramic coating, and C is the positive electrode coating area, which is the area on the positive electrode plate coated with the coating.

[0003] Currently, CCD vision inspection systems are used to inspect the quality of the cut electrodes. By calculating the color difference or grayscale difference between the active material coating (usually dark) and the ceramic coating (white), the normality of the coating area is determined.

[0004] In the process of developing this invention, the inventors discovered at least the following problems in the prior art: when the ceramic coating peels off and exposes the aluminum foil substrate, there is a significant color difference between the bright silver metallic color of the aluminum foil substrate and the dark electrode coating. This makes it impossible for the CCD visual inspection system to effectively distinguish between "normal ceramic coating-electrode color difference" and "abnormal exposed aluminum foil-electrode color difference," leading to misjudgment of coating defects and the identification of electrodes with serious safety hazards as qualified products. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide a detection system and method for defects in the ceramic coating of battery electrodes, which can improve the accuracy of identifying defects in the ceramic coating of battery electrodes.

[0006] A system for detecting defects in the ceramic coating of battery electrodes, comprising: Production line conveyor belt, rigid support, optical imaging unit, eddy current sensor, controller, and encoder mounted on the conveyor roller shaft. The production line conveyor belt transports the battery electrode sheet to be tested. The optical imaging unit and the eddy current sensor are mounted on the rigid bracket, and the acquisition area of ​​the optical imaging unit and the eddy current sensor corresponds to the detection area of ​​the battery electrode sheet to be tested. Using the synchronization pulse generated by the encoder, the optical imaging unit and the eddy current sensor are synchronously triggered to acquire data in the detection area of ​​the battery electrode to be tested, and obtain image signals and eddy current signals respectively. If the controller analyzes that the intensity of the eddy current signal exceeds the intensity of the signal from the uncoated aluminum foil, it determines that the coating in the detection area has peeled off. The controller inputs the image signal into the image recognition model to output a preliminary defect judgment result of the detection area of ​​the battery electrode to be detected; The controller uses the positive sample labels corresponding to the coating peeling in the detection area and the image feature vectors corresponding to the preliminary defect judgment results to train the image recognition model, and then uses the optimized image recognition model to detect defects in the ceramic coating of the battery electrode.

[0007] It also includes the light source; The light source is set in a side angle incident manner to enhance the visual contrast between the ceramic coating on the surface of the battery electrode and the substrate; The optical imaging unit includes a high-resolution line-scan CCD industrial camera or an area-array CCD industrial camera.

[0008] The field of view of the optical imaging unit and the detection point of the eddy current sensor are spatially aligned with the acquisition area; And / or, The eddy current sensor includes multiple sensor probes, which are used to perform multi-point parallel detection of the ceramic coating in the detection area.

[0009] The encoder identifies the operation time and location stamp using image signals and eddy current signals; The controller uses the operation time to perform spatiotemporal synchronization correlation between the image signal and the eddy current signal of the same detection area. After the coating in the detection area is delamination, the controller uses the operation time and the location stamp to determine the corresponding image signal. The image recognition model extracts features from the image signal. The features include one or more of the following: chromaticity value, grayscale statistical features, texture features, edge gradient features, and reflectivity features.

[0010] After the optimized and trained image recognition model detects the peeling of the ceramic coating on the battery electrode, the controller uses the eddy current signal of the detected battery electrode to verify the detection result of the image recognition model.

[0011] The controller is used to detect defects in the ceramic coating of battery electrodes by using the eddy current signal from the eddy current sensor and the optimized image recognition model according to preset weights. The optimized image recognition model outputs one or more of the following results: detection of ceramic coating peeling off of the battery electrode, detection of coating contamination, detection of coating scratches, and detection of foreign matter covering the coating.

[0012] The detection system also includes an audible and visual alarm and an actuator. The audible and visual alarm device, in response to the battery electrode ceramic coating defect sent by the controller, sends an alarm message and displays the battery electrode ceramic coating defect. The actuator, in response to a defect in the ceramic coating of the battery electrode sent by the controller, triggers a labeling operation and / or a rejection operation for the battery to be tested.

[0013] The controller constructs negative samples based on the signals corresponding to the normal coating in the detection area in the preliminary defect judgment result and the signals corresponding to the coating in the detection area in the preliminary defect judgment result. It also constructs positive samples based on the signals corresponding to the coating in the detection area in the preliminary defect judgment result and the signals corresponding to the coating in the detection area. The controller trains the image recognition model based on the negative samples and the positive samples. The training of the image recognition model includes online training and / or incremental training.

[0014] The image signals include one or more of the following: image signals acquired under different lighting conditions, image signals acquired under lens contamination conditions, and image signals of battery electrodes from different batches of electrode materials; The optimized image recognition model adapts to different lighting conditions, lens contamination, and battery electrodes corresponding to different batches of electrode materials.

[0015] According to a second aspect of the present invention, a method for detecting defects in the ceramic coating of a battery electrode is provided, comprising: The synchronous pulse generated by the encoder is used to synchronously trigger the optical imaging unit and the eddy current sensor to acquire data in the detection area of ​​the battery electrode to be tested, and obtain image signal and eddy current signal respectively. If the intensity of the eddy current signal exceeds the intensity of the uncoated aluminum foil signal, it is determined that the coating in the detection area has peeled off. The image signal is then input into an image recognition model to output a preliminary defect judgment result for the detection area of ​​the detected battery electrode. Using the positive sample labels corresponding to the coating peeling in the detection area and the preliminary defect judgment results, the image recognition model is trained, and the trained image recognition model is optimized to detect defects in the ceramic coating of the battery electrode.

[0016] According to a third aspect of the present invention, an electronic device for detecting defects in the ceramic coating of battery electrodes is provided, comprising: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors perform the methods described above.

[0017] According to a fourth aspect of the present invention, a computer-readable medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method as described above.

[0018] One embodiment of the above invention has the following advantages or beneficial effects: It utilizes an optical imaging unit and an eddy current sensor to acquire image signals and eddy current signals of the battery electrode to be tested, respectively, and trains an image recognition model based on the image signals and eddy current signals to identify defects in the ceramic coating of the battery electrode. Since the image recognition model is trained using various coating defects, it can improve the accuracy of identifying defects in the ceramic coating of the battery electrode.

[0019] The further effects of the aforementioned unconventional alternative methods will be explained below in conjunction with specific implementation methods. Attached Figure Description

[0020] The accompanying drawings are provided to better understand the invention and are not intended to unduly limit the scope of the invention. Wherein: Figure 1 This is a schematic diagram of the structure of the ceramic coating on the battery electrode. Figure 2 This is a schematic diagram of the main structure of a battery electrode ceramic coating defect detection system according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the main process of a method for detecting defects in the ceramic coating of battery electrodes according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of a computer system suitable for implementing terminal devices or servers of the present invention. Detailed Implementation

[0021] The following description, in conjunction with the accompanying drawings, illustrates exemplary embodiments of the present invention, including various details to aid understanding. These details should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the invention. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0022] The following methods are used in the existing technology to detect defects in the ceramic coating of battery electrodes.

[0023] Laser ranging / 3D vision solutions: Although they can detect the peeling of ceramic coatings on battery electrodes by detecting changes in thickness, the equipment is expensive, the data processing is complex, and the detection speed may be limited, making it difficult to popularize on high-speed production lines.

[0024] Single eddy current testing solution: It is very sensitive and reliable for exposed metal, but its function is limited. It cannot identify non-conductive defects on the coating (such as stains, damage but not complete exposure of aluminum, etc.) and cannot provide visual evidence, which is not conducive to quality traceability and analysis.

[0025] Therefore, detecting defects in the ceramic coating of battery electrodes presents a technical problem with a high error rate.

[0026] To improve the accuracy of identifying defects in the ceramic coating of battery electrodes, the following technical solutions from the embodiments of the present invention can be adopted.

[0027] See Figure 2 , Figure 2 This is a schematic diagram of the main structure of a battery electrode ceramic coating defect detection system according to an embodiment of the present invention. It includes: a production line conveyor belt 201, a rigid support 202, an optical imaging unit 203, an eddy current sensor 204, a controller 205, and an encoder 206 mounted on the conveyor roller shaft.

[0028] The production line conveyor belt 201 is horizontally positioned for continuously transporting battery electrode sheets to be tested along the testing direction. A rigid support 202 is fixedly installed at the testing station along the transport direction of the production line conveyor belt 201. The rigid support 202 spans across the production line conveyor belt 201.

[0029] Both the optical imaging unit 203 and the eddy current sensor 204 are fixedly mounted on the rigid bracket 202. The field of view center of the optical imaging unit 203 and the probe detection point of the eddy current sensor 204 coincide on the same detection axis of the production line conveyor belt 201 in the vertical projection direction, so as to ensure that the acquisition area of ​​the optical imaging unit 203 and the eddy current sensor 204 corresponds to the detection area of ​​the battery electrode to be tested.

[0030] The encoder 206 is coaxially mounted with the conveyor roller shaft of the production line conveyor belt 201. The signal output terminal of the encoder 206 is electrically connected to the input terminal of the optical imaging unit 203 and the input terminal of the eddy current sensor 204, respectively. When the conveyor roller shaft rotates, the encoder 206 generates a synchronous pulse signal proportional to the pole displacement in real time, and sends it to the optical imaging unit 203 and the eddy current sensor 204 simultaneously to achieve synchronous triggering.

[0031] The image signal output terminal of the optical imaging unit 203 is connected to the input terminal of the controller 205, and is used to transmit the acquired image signal to the controller 205. The eddy current signal output terminal of the eddy current sensor 204 is connected to the input terminal of the controller 205, and is used to transmit the eddy current signal to the controller 205.

[0032] The controller 205 receives and analyzes the eddy current signals collected by the eddy current sensor 204 in real time. The controller 205 presets the signal strength of the uncoated aluminum foil. Since the aluminum foil is a good conductor, when the ceramic coating peels off in the detection area of ​​the eddy current sensor 204, the electromagnetic coupling effect between the probe of the eddy current sensor 204 and the exposed aluminum foil is enhanced, resulting in a significant change in the eddy current signal strength.

[0033] When the controller 205 analyzes and finds that the intensity of the eddy current signal exceeds the intensity of the uncoated aluminum foil signal, it directly determines that there is a coating peeling defect in the detection area, that is, the coating in the detection area has peeled off.

[0034] Simultaneously, the controller 205 processes the image signals synchronously acquired by the optical imaging unit 203. Specifically, the controller 205 calls an image recognition model to extract and analyze features of the detection area in the image signal. Based on the image recognition model, the controller 205 outputs a preliminary defect judgment result for the detection area of ​​the battery electrode to be inspected. For example, it may determine that the coating has peeled off. It should be noted that this preliminary defect judgment result is uncertain because the image recognition model has not been fully trained.

[0035] The controller 205 trains the image recognition model by using the positive sample labels corresponding to coating peeling in the detection area and the image feature vectors corresponding to the preliminary defect judgment results output by the image recognition model. After the image recognition model completes optimized training, it can directly detect defects in the ceramic coating of the battery electrode.

[0036] In the above embodiment, during the transmission of the battery electrode to be inspected on the production line conveyor belt 201, an image recognition model is trained by using an optical imaging unit 203 and an eddy current sensor 204 mounted on a rigid support 202 to detect the battery electrode. This image recognition model is then used to accurately identify defects in the ceramic coating of the battery electrode.

[0037] In one embodiment of the present invention, if the intensity of the eddy current signal collected by the eddy current sensor 204 exceeds the signal intensity of the uncoated aluminum foil, the controller 205 directly determines that the coating of the battery electrode to be inspected has peeled off in the inspection area. That is, based on the production line conveyor belt 201, rigid support 202, eddy current sensor 204, and encoder 206, defects in the ceramic coating of the battery electrode can be directly detected. In one embodiment of the present invention, the optical imaging unit 203 uses a high-resolution line-scanning CCD industrial camera. Considering that the battery electrode is in a continuous high-speed motion state on the production line, the high-resolution line-scanning CCD industrial camera can stitch together a complete two-dimensional image of the battery electrode surface without distortion by scanning line by line, making it particularly suitable for inspection scenarios with a wide area and high movement speed.

[0038] In addition, for production lines with specific detection accuracy requirements or intermittent transmission, the optical imaging unit 203 can adopt an area-array CCD industrial camera. The area-array CCD industrial camera acquires two-dimensional image information, and with the appropriate electronic shutter, it can effectively acquire images of moving electrode sheets.

[0039] Regardless of whether a high-resolution line-scan CCD industrial camera or an area-scan CCD industrial camera is used, the resolution of the optical imaging unit 203 needs to ensure that it can clearly identify minute ceramic coating peeling defects and surface foreign objects.

[0040] The system for detecting defects in the ceramic coating of battery electrodes also includes a light source 207. The light source 207 is positioned with a side angle incidence to enhance the visual contrast between the ceramic coating on the battery electrode surface and the substrate.

[0041] The ceramic coating on the surface of the battery electrode is a diffuse reflective material, producing uniform anisotropic scattering of incident light. The exposed aluminum foil substrate has specular reflective properties, producing directional reflection of incident light. When conventional vertical illumination is used, the difference in photosensitive intensity between the two on the optical imaging unit 203 is not significant.

[0042] When using a side-angle incident light method: for the flat ceramic coating area, the light is uniformly diffused, and some of the light enters the optical imaging unit 203, presenting grayscale information. For the exposed aluminum foil area after the coating has peeled off, the side-incident light undergoes specular reflection on the smooth aluminum foil surface, and the reflected light deviates from the optical imaging unit 203, causing this area to appear significantly darker in the image. By utilizing the difference in reflective properties between the ceramic coating and the aluminum foil substrate, the visual contrast in the image is amplified by setting the incident angle of the light source 207.

[0043] In one embodiment of the present invention, to ensure effective fusion of multi-source heterogeneous data, the field of view of the optical imaging unit 203 and the detection point of the eddy current sensor 204 are spatially aligned with the acquisition area. For example, the mounting bases of the optical imaging unit 203 and the eddy current sensor 204 in the rigid bracket 202 are adjusted to achieve spatial alignment of the field of view of the optical imaging unit 203 and the detection point of the eddy current sensor 204 with the acquisition area. This ensures the accuracy of the mapping from the eddy current signal to the image signal and avoids model mistraining caused by misalignment of the acquisition area.

[0044] Given the wide surface area of ​​battery electrodes and the fact that the tab area may be distributed on both sides of the electrode, a single eddy current sensor 204 probe is insufficient to cover the entire detection area.

[0045] The eddy current sensor 204 includes multiple sensor probes, which are used to perform multi-point parallel detection of the ceramic coating in the detection area.

[0046] Multiple sensor probes are arranged on a rigid support 202 along a width direction perpendicular to the direction of battery electrode transport. The detection points of each probe are connected to form a transverse detection line, which coincides spatially with the line scan field of view of the optical imaging unit 203.

[0047] During the inspection process, multiple sensor probes are triggered by synchronous pulse signals generated by encoder 206, and acquire their respective eddy current signals in parallel. Controller 205 receives and processes the multiple eddy current signals from each probe, and independently determines whether coating peeling defects exist at each lateral position.

[0048] In the above embodiments, the eddy current sensor 204 includes multiple sensor probes, eliminating the detection blind zone caused by the width of the battery electrode being larger than the detection range of a single sensor. Furthermore, it can accurately locate the specific position of defects in the electrode width direction.

[0049] In one embodiment of the present invention, the controller 205 spatiotemporally correlates the image signal and eddy current signal of the same detection area through the operation time. After the coating of the detection area is determined to be peeled off, the corresponding image signal is determined by the operation time and the location stamp. The image recognition model extracts features from the image signal.

[0050] Specifically, a unified data spatiotemporal coordinate system is constructed within the controller 205. In this coordinate system, each eddy current data point acquired by the eddy current sensor 204, and each line of image data acquired by the optical imaging unit 203, is assigned unique identification information, including a timestamp and a location stamp. The operation time can be determined based on the timestamp.

[0051] When the controller 205 analyzes the timestamp T and location stamp P corresponding to the detection point, and the eddy current signal intensity of the detection point exceeds the signal intensity of the uncoated aluminum foil, confirming that the coating in the detection area has peeled off, the controller 205 immediately performs a reverse tracing operation.

[0052] The reverse tracing operation includes retrieving image signals acquired by the optical imaging unit 203 at the same time and location, using timestamp T and location stamp P as indexes. Since the optical imaging unit 203 and the eddy current sensor 204 are synchronously triggered by the same encoder pulse, and their detection areas are pre-aligned in space, this retrieval operation can accurately lock the image signals.

[0053] The controller 205 employs an image recognition model to extract features from the image signal, constructing a feature vector that comprehensively describes the visual characteristics of the defects. These features include one or more of the following: chromaticity values, grayscale statistical features, texture features, edge gradient features, and reflectivity features. Specifically, the reflectivity feature describes the reflectivity of the battery electrodes under illumination conditions using light sources 207 at different angles.

[0054] In one embodiment of the present invention, the optimized and trained image recognition model already possesses strong detection capabilities, enabling it to independently identify ceramic coating peeling defects with high accuracy. However, considering the extremely stringent safety requirements of battery electrodes as core components of lithium batteries, further verification based on eddy current signals is performed to improve detection reliability.

[0055] The optimized image recognition model in controller 205 processes the image signals acquired in real time by optical imaging unit 203 and outputs the defect type and confidence score for each detection area. When a detection area is identified as coating peeling and the confidence score is higher than the preset image threshold, the defect judgment result can be directly output.

[0056] When a detection area is identified as having coating peeling and the confidence level is lower than or equal to a preset image threshold, the controller 205 uses the operation time and location stamp corresponding to the suspected defect area output by the image recognition model as an index to query the eddy current signal of that detection area. If the intensity of the eddy current signal exceeds the intensity of the uncoated aluminum foil signal, the coating peeling in that detection area is confirmed, meaning the detection result of the image recognition model is correct. If the intensity of the eddy current signal does not exceed the intensity of the uncoated aluminum foil signal, it indicates that there is still a ceramic coating covering the detection area, and the image recognition model may generate a false alarm. The controller 205 determines that the detection area is normal.

[0057] In the above embodiments, eddy current signals are used for verification to ensure the absolute reliability of high-risk defect detection.

[0058] In one embodiment of the present invention, two independent yet cooperative detection channels are formed: one is an eddy current signal detection channel, and the other is an image recognition model channel.

[0059] The controller 205 uses preset weights to fuse the two detection channels. That is, it uses the eddy current signal from the eddy current sensor 204 and the optimized image recognition model according to preset weights to detect defects in the ceramic coating of the battery electrode.

[0060] For example, in order to detect the ceramic coating peeling detection results, the preset weight of the eddy current signal of the eddy current sensor 204 is 0.8, and the preset weight of the optimized image recognition model is 0.2.

[0061] In order to detect coating contamination, coating scratches, or foreign matter on the coating, the preset weight of the eddy current signal of the eddy current sensor 204 is 0.1, and the preset weight of the optimized image recognition model is 0.9.

[0062] In the above embodiments, the controller 205 uses preset weights to fully detect the battery electrodes in both channels for different detection purposes, thereby improving detection accuracy.

[0063] In one embodiment of the present invention, the image recognition model is trained using positive and negative samples. Positive samples refer to image regions that can characterize coating peeling and their corresponding visual feature vectors. Negative samples refer to image regions that can characterize normal coating and their corresponding visual feature vectors. Both positive and negative samples are constructed by the controller 205.

[0064] When the eddy current signal indicates coating peeling, and the image recognition model's preliminary defect assessment of the area also indicates coating peeling, the sample is labeled as a positive sample. When the eddy current signal indicates coating peeling, and the image recognition model's preliminary defect assessment of the area indicates the coating is normal, the sample is labeled as a negative sample.

[0065] The image recognition model is trained using the aforementioned negative and positive samples. Training methods include online training and / or incremental training. For example, the controller 205 has a cache queue. The cache queue is used to temporarily store positive and negative samples acquired in real time and subjected to spatiotemporal correlation. When the number of samples in the cache queue reaches a preset number, an online training session is triggered. Compared to online training, incremental training uses longer time intervals or a larger accumulation of samples as triggering conditions, performing more thorough and in-depth training.

[0066] In one embodiment of the present invention, in a real production environment, the detection of battery electrodes faces various dynamically changing interference factors. If the image recognition model does not fully learn these interference factors, it will lead to insufficient generalization ability and an increased false detection rate. Therefore, the image signals include signals from various operating conditions: image signals acquired under different lighting conditions, image signals acquired under lens contamination conditions, and image signals from battery electrodes made of different batches of electrode materials. The optimized and trained image recognition model adapts to different lighting conditions, lens contamination, and battery electrodes from different batches of electrode materials.

[0067] In one embodiment of the invention, the detection system further includes an audible and visual alarm and an actuator. The audible and visual alarm is installed within the operator's visual and audible range near the detection station. In response to a defect in the battery electrode ceramic coating sent by the controller 205, the audible and visual alarm sends an alarm message and displays the defect. For example, the defect in the battery electrode ceramic coating can be displayed on a human-machine interface screen.

[0068] The actuator, in response to a defect in the ceramic coating of the battery electrode sent by the controller 205, triggers a labeling operation and / or rejection operation for the battery to be tested.

[0069] See Figure 3 , Figure 3 This is a schematic diagram of the main flow of a method for detecting defects in the ceramic coating of battery electrodes according to an embodiment of the present invention, specifically including the following steps: S301. Using the synchronous pulse generated by the encoder, the optical imaging unit and the eddy current sensor are synchronously triggered to acquire data in the detection area of ​​the battery electrode to be tested, and obtain image signals and eddy current signals respectively.

[0070] In one embodiment of the present invention, the light source is arranged in a side angle incident manner to enhance the visual contrast between the ceramic coating on the surface of the battery electrode and the substrate. The optical imaging unit includes a high-resolution line-scan CCD industrial camera or an area-array CCD industrial camera.

[0071] In one embodiment of the present invention, the field of view of the optical imaging unit and the detection point of the eddy current sensor are spatially aligned with the acquisition area; And / or, The eddy current sensor includes multiple sensor probes, which are used to perform multi-point parallel detection of the ceramic coating in the detection area.

[0072] In one embodiment of the present invention, the encoder identifies the operation time and location stamp using image signals and eddy current signals; The controller uses the operation time to perform spatiotemporal synchronization correlation between the image signal and the eddy current signal of the same detection area. After the coating in the detection area is delamination, the controller uses the operation time and the location stamp to determine the corresponding image signal. The image recognition model extracts features from the image signal. The features include one or more of the following: chromaticity value, grayscale statistical features, texture features, edge gradient features, and reflectivity features.

[0073] S302. If the intensity of the eddy current signal exceeds the intensity of the uncoated aluminum foil signal, it is determined that the coating of the detection area of ​​the detected battery electrode has peeled off. The image signal is then input into the image recognition model to output the preliminary defect judgment result of the detection area.

[0074] S303. Using the positive sample labels corresponding to the coating peeling in the detection area and the image feature vectors corresponding to the preliminary defect judgment results, train the image recognition model, and detect defects in the ceramic coating of the battery electrode by optimizing the trained image recognition model.

[0075] In one embodiment of the present invention, after the optimized and trained image recognition model detects the peeling of the ceramic coating on the battery electrode, the controller uses the eddy current signal of the detected battery electrode to verify the detection result of the image recognition model.

[0076] In one embodiment of the present invention, the controller is used to detect defects in the ceramic coating of battery electrode sheets by using the eddy current signal of the eddy current sensor and the optimized image recognition model according to preset weights. The optimized image recognition model outputs one or more of the following results: detection of ceramic coating peeling off of the battery electrode, detection of coating contamination, detection of coating scratches, and detection of foreign matter covering the coating.

[0077] In one embodiment of the present invention, the detection system further includes an audible and visual alarm and an actuator. The audible and visual alarm device, in response to the battery electrode ceramic coating defect sent by the controller, sends an alarm message and displays the battery electrode ceramic coating defect. The actuator, in response to a defect in the ceramic coating of the battery electrode sent by the controller, triggers a labeling operation and / or a rejection operation for the battery to be tested.

[0078] In one embodiment of the present invention, the controller constructs negative samples based on the signals corresponding to the normal coating in the detection area in the preliminary defect judgment result and the signals corresponding to the coating in the detection area in the preliminary defect judgment result, and constructs positive samples based on the signals corresponding to the coating in the detection area in the preliminary defect judgment result. The image recognition model is trained according to the negative samples and the positive samples. The training of the image recognition model includes online training and / or incremental training.

[0079] In one embodiment of the present invention, the image signal includes one or more of the following: image signals acquired under different lighting conditions, image signals acquired under lens contamination conditions, and image signals of battery electrodes from different batches of electrode materials; The optimized image recognition model adapts to different lighting conditions, lens contamination, and battery electrodes corresponding to different batches of electrode materials.

[0080] The following is for reference. Figure 4 It shows a schematic diagram of the structure of a computer system 400 suitable for implementing a terminal device of the present invention. Figure 4 The terminal device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0081] like Figure 4As shown, the computer system 400 includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 402 or programs loaded from storage section 408 into random access memory (RAM) 403. The RAM 403 also stores various programs and data required for the operation of the system 400. The CPU 401, ROM 402, and RAM 403 are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0082] The following components are connected to I / O interface 405: an input section 406 including a keyboard, mouse, etc.; an output section 407 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN card, modem, etc. The communication section 409 performs communication processing via a network such as the Internet. Drive 410 is also connected to I / O interface 405 as needed. Removable media 411, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 410 as needed so that computer programs read from them can be installed into storage section 408 as needed.

[0083] In particular, according to the embodiments disclosed in this invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by central processing unit (CPU) 401, it performs the functions defined above in the system of this invention.

[0084] It should be noted that the computer-readable medium shown in this invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0085] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0086] The modules described in the embodiments of the present invention can be implemented in software or hardware. The described modules can also be housed in a processor. In another aspect, the present invention also provides a computer-readable medium, which may be included in the device described in the above embodiments; or it may exist independently and not assembled into the device. The computer-readable medium carries one or more programs, which, when executed by the device, cause the device to include: The synchronous pulse generated by the encoder is used to synchronously trigger the optical imaging unit and the eddy current sensor to acquire data in the detection area of ​​the battery electrode to be tested, and obtain image signal and eddy current signal respectively. If the intensity of the eddy current signal exceeds the intensity of the uncoated aluminum foil signal, it is determined that the coating in the detection area has peeled off. The image signal is then input into an image recognition model to output a preliminary defect judgment result for the detection area of ​​the detected battery electrode. The image recognition model is trained by using the positive sample labels corresponding to the coating peeling in the detection area and the image feature vectors corresponding to the preliminary defect judgment results. The trained image recognition model is then used to detect defects in the ceramic coating of the battery electrode.

[0087] According to the technical solution of this invention, an optical imaging unit and an eddy current sensor are used to acquire image signals and eddy current signals of the battery electrode to be tested, respectively. An image recognition model is trained based on the image signals and eddy current signals to identify defects in the ceramic coating of the battery electrode. Since the image recognition model is trained using various coating defects, the accuracy of identifying defects in the ceramic coating of the battery electrode can be improved.

[0088] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention. It should be noted that the acquisition, storage, and application of user personal information involved in the technical solutions of this disclosure comply with relevant laws and regulations and do not violate public order and good morals.

Claims

1. A detection system for defects in the ceramic coating of battery electrodes, characterized in that, include: Production line conveyor belt, rigid support, optical imaging unit, eddy current sensor, controller, and encoder mounted on the conveyor roller shaft. The production line conveyor belt transports the battery electrode sheet to be tested. The optical imaging unit and the eddy current sensor are mounted on the rigid bracket, and the acquisition area of ​​the optical imaging unit and the eddy current sensor corresponds to the detection area of ​​the battery electrode sheet to be tested. Using the synchronization pulse generated by the encoder, the optical imaging unit and the eddy current sensor are synchronously triggered to acquire data in the detection area of ​​the battery electrode to be tested, and obtain image signals and eddy current signals respectively. If the controller analyzes that the intensity of the eddy current signal exceeds the intensity of the signal from the uncoated aluminum foil, it determines that the coating in the detection area has peeled off. The controller inputs the image signal into the image recognition model to output a preliminary defect judgment result of the detection area of ​​the battery electrode to be detected; The controller uses the positive sample labels corresponding to the coating peeling in the detection area and the image feature vectors corresponding to the preliminary defect judgment results to train the image recognition model, and then uses the optimized image recognition model to detect defects in the ceramic coating of the battery electrode.

2. The detection system for defects in the ceramic coating of battery electrodes according to claim 1, characterized in that, It also includes the light source; The light source is set in a side angle incident manner to enhance the visual contrast between the ceramic coating on the surface of the battery electrode and the substrate; The optical imaging unit includes a high-resolution line-scan CCD industrial camera or an area-array CCD industrial camera.

3. The detection system for defects in the ceramic coating of battery electrodes according to claim 1, characterized in that, The field of view of the optical imaging unit and the detection point of the eddy current sensor are spatially aligned with the acquisition area; And / or, The eddy current sensor includes multiple sensor probes, which are used to perform multi-point parallel detection of the ceramic coating in the detection area.

4. The detection system for defects in the ceramic coating of battery electrodes according to claim 1, characterized in that, The encoder identifies the operation time and location stamp using image signals and eddy current signals; The controller uses the operation time to perform spatiotemporal synchronization correlation between the image signal and the eddy current signal of the same detection area. After the coating in the detection area is delamination, the controller uses the operation time and the location stamp to determine the corresponding image signal. The image recognition model extracts features from the image signal. The features include one or more of the following: chromaticity value, grayscale statistical features, texture features, edge gradient features, and reflectivity features.

5. The detection system for defects in the ceramic coating of battery electrodes according to claim 1, characterized in that, After the optimized and trained image recognition model detects the peeling of the ceramic coating on the battery electrode, the controller uses the eddy current signal of the detected battery electrode to verify the detection result of the image recognition model.

6. The detection system for defects in the ceramic coating of battery electrodes according to claim 1, characterized in that, The controller is used to detect defects in the ceramic coating of battery electrodes by using the eddy current signal from the eddy current sensor and the optimized image recognition model according to preset weights. The optimized image recognition model outputs one or more of the following results: detection of ceramic coating peeling off of the battery electrode, detection of coating contamination, detection of coating scratches, and detection of foreign matter covering the coating.

7. The detection system for defects in the ceramic coating of battery electrodes according to claim 1, characterized in that, The detection system also includes an audible and visual alarm and an actuator. The audible and visual alarm device, in response to the battery electrode ceramic coating defect sent by the controller, sends an alarm message and displays the battery electrode ceramic coating defect. The actuator, in response to a defect in the ceramic coating of the battery electrode sent by the controller, triggers a labeling operation and / or a rejection operation for the battery to be tested.

8. The detection system for defects in the ceramic coating of battery electrodes according to claim 1, characterized in that, The controller constructs negative samples based on the signals corresponding to the normal coating in the detection area in the preliminary defect judgment result and the signals corresponding to the coating in the detection area in the preliminary defect judgment result. It also constructs positive samples based on the signals corresponding to the coating in the detection area in the preliminary defect judgment result and the signals corresponding to the coating in the detection area. The controller trains the image recognition model based on the negative samples and the positive samples. The training of the image recognition model includes online training and / or incremental training.

9. The detection system for defects in the ceramic coating of battery electrodes according to claim 1, characterized in that, The image signals include one or more of the following: image signals acquired under different lighting conditions, image signals acquired under lens contamination conditions, and image signals of battery electrodes from different batches of electrode materials; The optimized image recognition model adapts to different lighting conditions, lens contamination, and battery electrodes corresponding to different batches of electrode materials.

10. A method for detecting defects in the ceramic coating of battery electrodes, characterized in that, Applied to the system of claim 1, comprising: The synchronous pulse generated by the encoder is used to synchronously trigger the optical imaging unit and the eddy current sensor to acquire data in the detection area of ​​the battery electrode to be tested, and obtain image signal and eddy current signal respectively. If the intensity of the eddy current signal exceeds the intensity of the uncoated aluminum foil signal, it is determined that the coating in the detection area has peeled off. The image signal is then input into an image recognition model to output a preliminary defect judgment result for the detection area of ​​the detected battery electrode. The image recognition model is trained by using the positive sample labels corresponding to the coating peeling in the detection area and the image feature vectors corresponding to the preliminary defect judgment results. The trained image recognition model is then used to detect defects in the ceramic coating of the battery electrode.

Images