New energy battery module appearance detection system based on machine vision

Abstract

The application discloses a new energy battery module appearance detection system based on machine vision and relates to the field of visual detection, which comprises a camera module, a storage module and the like.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of visual detection, in particular to a new energy battery module appearance detection system based on machine vision. BACKGROUND

[0002] The new energy battery module is the core energy storage unit of the new energy equipment, which is composed of the battery cell, the BMS, the structure and the heat dissipation part.

[0003] The application number 202510383050.0 discloses a new energy lithium battery surface defect visual detection method based on Y0L0v8 model, which comprises: constructing a multi-modal optical imaging image acquisition system, optimizing the optical parameters of the image acquisition system; collecting lithium ion battery surface defect pictures based on the optimized image acquisition system, constructing a standardized data set based on the lithium ion battery surface defect pictures; constructing an improved Y0L0v8 network based on the CSPDarknet53 structure and the Anchor-free detection head, training the improved Y0L0v8 network based on the standardized data set to obtain a deep learning detection engine; detecting the surface defects of the new energy lithium battery based on the deep learning detection engine, which aims to solve the problem that the current lithium battery manufacturing field generally adopts a quality control scheme combining manual visual inspection and semi-automatic detection, which has significant technical limitations: firstly, the detection efficiency problem, the manual detection speed is limited, which is seriously mismatched with the walking speed of the modern production line of >30m / min. Secondly, manual visual inspection will have subjective judgment deviation: the experience difference of the detection personnel and individual visual fatigue will cause the defect recognition consistency coefficient to be less than 0.65, and the false detection rate is as high as 15%-20%. In addition, due to the lack of systematic defect feature data set, it is difficult to realize the correlation analysis of process parameters and quality defects.

[0004] However, the prior art has limited precision in detecting the appearance quality of the new energy battery module due to the limited number of appearance images of the new energy battery module collected.

[0005] Therefore, we propose a new energy battery module appearance detection system based on machine vision. SUMMARY

[0006] In view of the above shortcomings of the prior art, the present application provides a new energy battery module appearance detection system based on machine vision, which can effectively solve the problems of the prior art.

[0007] To achieve the above purpose, the present application realizes the following technical scheme;

[0008] The present application discloses a new energy battery module appearance detection system based on machine vision, which comprises:

[0009] The camera module is used to acquire surface images of the battery module and preprocess these images to extract the corresponding contour images. The storage module stores standard battery module image data and receives the contour images output by the camera module, temporarily storing the contour images. The matching module retrieves the contour images stored in the storage module and the temporarily stored contour images, matches the two sets of contour images, and then forwards the matched contour images to the analysis module for analysis. The analysis module receives the matched contour images, evaluates their similarity, and analyzes whether the battery module's appearance is acceptable. The output and interaction module receives the analysis results from the analysis module, sorts the currently inspected battery modules based on the analysis results, and outputs the analysis results to the control panel of the battery module production equipment.

[0010] The camera module is interconnected with a storage module via a wireless network. The storage module is interconnected with a matching module and an analysis module via a wireless network. The matching module is interconnected with an identification unit via a wireless network. The analysis module is interconnected with an output and interaction module via a wireless network.

[0011] Furthermore, the camera module includes a slider, a slide block, a slip ring, and a high-speed high-definition camera;

[0012] The battery module is output from its production equipment and is conveyed by the conveyor belt assembly of the production equipment to the area directly below the slip ring, so that the center point of the plane of the slip ring in a horizontal state is on the same vertical line as the center point of the top surface of the battery module directly below it.

[0013] The slide bar is fixedly connected to the ground and located on one side of the production equipment conveyor belt assembly. A slider is installed on the side of the slide bar near the conveyor belt assembly via an electric slide rail. The end of the slider away from the slide bar is connected to a slip ring via an electric rotating shaft. A high-speed high-definition camera is installed on the inner ring wall of the slip ring via an electric slide rail. During the initial operation phase of the system, the slider takes the upper limit position of the electric slide rail on the slide bar as the initial position. When the slip ring is carried by the slider to the initial position, the slip ring is in a horizontal posture.

[0014] In this process, the battery modules output by the generating device are in the same orientation relative to the surface of the conveyor belt assembly each time.

[0015] During the initial operation phase of the system, a standard battery module is placed on the conveyor belt assembly of the production equipment and transported by the conveyor belt assembly to the area directly below the slip ring. The rotation angle is set to drive the electric shaft to rotate the slip ring to the target angle. The electric slide rail on the drive slide rod pushes the slider to carry the slip ring downward until the slip ring surrounds the standard battery module. Due to the rotation into an inclined posture, the highest point of the slip ring surface is not higher than the top surface of the battery module, and the lowest point is not lower than the surface of the conveyor belt assembly of the production equipment.

[0016] A high-speed, high-definition camera moves in a circular motion at a predetermined speed in an electric slide rail on the inner wall of the slip ring, and continuously captures images of the battery module surface at a predetermined shooting rate. The captured images of the battery module surface are simultaneously preprocessed to extract contour images. The resulting set of contour images is the standard battery module image data stored in the storage module.

[0017] The battery modules to be tested output by the conveyor belt assembly of the production equipment perform the same operation as above to obtain their corresponding contour image set, and send them to the storage module at the same time. The images are temporarily stored in the storage module, and the temporarily stored contour image set is cleared after the system obtains the result of whether it is qualified.

[0018] Furthermore, when the high-speed high-definition camera in the camera module acquires surface images of different battery modules, the initial acquisition position of each battery module is consistent with the slip ring where the high-speed high-definition camera is located during the acquisition operation of different battery modules. The high-speed high-definition camera has a built-in position sensor, and the operating frequency of the position sensor is consistent with the shooting rate of the high-speed high-definition camera. After each battery module surface image is captured, its positioning result at the same time is applied for marking. After the contour image is extracted from the battery module surface image, the contour image uses the marking of its source battery module surface image.

[0019] The standard battery modules transported by the conveyor belt assembly of the production equipment perform the same operations as above during the acquisition of the contour image set.

[0020] Among them, the operating frequency of the sensor, the shooting rate of the high-speed high-definition camera, and the predetermined speed of the high-speed high-definition camera in the electric slide rail on the inner wall of the slip ring follow the following rules: the higher the detection accuracy requirement, the faster the operating frequency and shooting rate, and the slower the predetermined speed; conversely, the slower the operating frequency and shooting rate, the faster the predetermined speed.

[0021] Furthermore, the preprocessing of the surface image of the battery module includes:

[0022] ;

[0023] In the formula: This represents the pixel value at position (x, y) in the preprocessed image. As a key intermediate variable; This represents the pixel value at position (x, y) in the original image; This is the numerical stability constant, with a value of 1;

[0024] In this process, each pixel in the surface image of the battery module is processed based on the above formula to output a preprocessed surface image of the battery module, and a contour image is extracted based on the preprocessed surface image of the battery module.

[0025] Furthermore, the aforementioned The answer is obtained using the following formula:

[0026] ;

[0027] In the formula: For neighborhood size parameters, ; Normalized Gaussian kernel; Fill the boundary with pixels;

[0028] in, , Let (u, v) represent the Gaussian kernel normalization factor, and (u, v) represent the relative coordinates of the neighborhood centered at (x, y). Both u and v are within the range of (x, y). Within the range, This represents the Gaussian kernel standard deviation. ∈[0.5,2.0];

[0029] in, , This represents the actual width and height of the image. When p < 0, =0, when p≥W When q < 0, =0, when , .

[0030] Furthermore, the matching module is equipped with a recognition unit, which is used to recognize the marking information of all contour images acquired by the matching module, so that contour images with consistent marking information can be matched with each other. That is, each set of matched contour images contains two contour images, which are respectively from the standard battery module image data set stored in the storage module and the temporarily stored contour image set.

[0031] Furthermore, during the operation of the analysis module, when analyzing whether the appearance of the battery module is qualified, the comprehensive similarity evaluation result of each matched contour image is used as the quantitative standard for whether the analysis is qualified:

[0032] ;

[0033] In the formula: For comprehensive similarity; Number of groups to match for the contour image; Let be the similarity of the r-th group of contour images; Configure weights for the r-th group of contour images; These are the curvature values ​​at the positions on contour a and contour b, respectively, at the positions on contour b and contour b, at the positions on contour b and contour b, respectively, at the positions on contour a .... This means taking the shortest contour line between the two contours for integration calculation; The bounded area of ​​the intersection region of the two contours; The bounded area of ​​the region formed by the union of the two contours; Let be the perimeter of the intersection region of the two contours; Let be the perimeter of the outer contour of the region where the two contours meet; ∮ is the curve integral along the contour.

[0034] in, During computation, the center points of the two matched contour images are aligned before the computation operation is performed. After obtaining the result, it is compared with the preset qualified threshold in the analysis module. If the battery module's appearance is not less than the preset acceptable threshold, it is considered acceptable; otherwise, it is considered unacceptable.

[0035] Furthermore, the aforementioned The value follows:

[0036] Calculate the mean foreground-to-background ratio of each matched contour image in its source image, denoted as . ,but , express The summation result.

[0037] Compared with the known prior art, the technical solution provided by this invention has the following beneficial effects:

[0038] This invention provides a machine vision-based appearance inspection system for new energy battery modules. During operation, the system acquires surface images of battery modules through specific methods and optimizes preprocessing to efficiently extract contour images, ensuring the integrity and accuracy of image information. Furthermore, through a consistent acquisition, positioning, and marking mechanism, the system achieves more precise matching of contour images from different battery modules, reducing matching errors. Based on multi-dimensional parameters, the system calculates comprehensive similarity and, combined with preset thresholds, scientifically determines whether the appearance is acceptable, improving the reliability of inspection results. Temporary image data is promptly cleared after inspection to avoid redundant data consuming resources. Simultaneously, the system can flexibly adjust operating parameters according to inspection accuracy requirements to adapt to different inspection scenarios, effectively improving the efficiency and quality of battery module appearance inspection, assisting production equipment in achieving precise sorting, and ensuring stable battery module production quality. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0040] Figure 1 This is a schematic diagram of a machine vision-based new energy battery module appearance inspection system.

[0041] Figure 2 This is a schematic diagram of the camera module structure in this invention.

[0042] The numbers in the diagram represent: 1. Sliding rod; 2. Sliding block; 3. Slip ring; 4. High-speed high-definition camera. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0044] The present invention will be further described below with reference to embodiments.

[0045] Example:

[0046] The machine vision-based new energy battery module appearance inspection system in this embodiment, such as... Figure 1 As shown, it includes:

[0047] The camera module is used to acquire surface images of the battery module and preprocess the surface images of the battery module to extract the contour images corresponding to the surface images of the battery module.

[0048] The camera module includes a slider, a slide ring, and a high-speed HD camera;

[0049] The battery module is output from its production equipment and transported by the conveyor belt assembly of the production equipment to the area directly below the slip ring, so that the center point of the plane of the slip ring in a horizontal state is on the same vertical line as the center point of the top surface of the battery module directly below it.

[0050] The slide bar is fixedly connected to the ground and located on one side of the production equipment conveyor belt assembly. A slider is installed on the side of the slide bar closest to the conveyor belt assembly via an electric slide rail. The end of the slider away from the slide bar is connected to a slip ring via an electric rotating shaft. A high-speed high-definition camera is installed on the inner ring wall of the slip ring via an electric slide rail. During the initial operation of the system, the slider takes the upper limit position of the electric slide rail on the slide bar as the initial position. When the slip ring is in the initial position carried by the slider, the slip ring is in a horizontal posture.

[0051] In this process, the battery modules output by the generating device are in the same orientation relative to the surface of the conveyor belt assembly each time.

[0052] During the initial operation phase of the system, a standard battery module is placed on the conveyor belt assembly of the production equipment and transported by the conveyor belt assembly to the area directly below the slip ring. The rotation angle is set to drive the electric shaft to rotate the slip ring to the target angle. The electric slide rail on the drive slide rod pushes the slider to carry the slip ring downward until the slip ring surrounds the standard battery module. Due to the rotation into an inclined posture, the highest point of the slip ring surface is not higher than the top surface of the battery module, and the lowest point is not lower than the surface of the conveyor belt assembly of the production equipment.

[0053] A high-speed, high-definition camera moves in a circular motion at a predetermined speed in an electric slide rail on the inner wall of the slip ring, and continuously captures images of the battery module surface at a predetermined shooting rate. The captured images of the battery module surface are simultaneously preprocessed to extract contour images. The resulting set of contour images is the standard battery module image data stored in the storage module.

[0054] The battery modules to be tested output by the production equipment conveyor belt assembly perform the same operation as above to obtain their corresponding contour image set, and send them to the storage module at the same time. They are temporarily stored in the storage module, and after the system obtains the result of whether they are qualified, the temporarily stored contour image set is cleared.

[0055] When the high-speed high-definition camera in the camera module acquires surface images of different battery modules, the initial acquisition position of each battery module is consistent with the slip ring where the high-speed high-definition camera is located during the acquisition operation of different battery modules. The high-speed high-definition camera has a built-in position sensor, and the operating frequency of the position sensor is consistent with the shooting rate of the high-speed high-definition camera. After each battery module surface image is captured, its positioning result at the same time is applied for marking. After the contour image is extracted from the battery module surface image, the contour image uses the marking of its source battery module surface image.

[0056] The standard battery modules transported by the production equipment conveyor belt assembly perform the same operations as above during the acquisition of the contour image set;

[0057] Among them, the operating frequency of the sensor, the shooting rate of the high-speed high-definition camera, and the predetermined speed of the high-speed high-definition camera in the electric slide rail on the inner wall of the slip ring follow the following rules: the higher the detection accuracy requirement, the faster the operating frequency and shooting rate, and the slower the predetermined speed; conversely, the slower the operating frequency and shooting rate, the faster the predetermined speed.

[0058] Preprocessing of the surface image of the battery module includes:

[0059] ;

[0060] In the formula: This represents the pixel value at position (x, y) in the preprocessed image. As a key intermediate variable; This represents the pixel value at position (x, y) in the original image; This is the numerical stability constant, with a value of 1;

[0061] The above formula calculates the pixel value of the preprocessed image by introducing key intermediate variables and numerical stability constants for each pixel in the original image. The setting of numerical stability constants can effectively avoid calculation anomalies caused by fluctuations in the original pixel values ​​or special values ​​(such as pixel values ​​close to 0) during the calculation process, ensuring the stability of the preprocessing results. The introduction of key intermediate variables can provide a more accurate pixel basis for subsequent contour image extraction based on the preprocessed image, solving the problem that the contour extraction is easily interfered with by factors such as noise and uneven lighting when directly using the original image. This makes the contour extraction more consistent with the actual shape of the battery module surface, laying the foundation for the accuracy of subsequent appearance inspection.

[0062] In this process, each pixel in the surface image of the battery module is processed based on the above formula to output a preprocessed surface image of the battery module, and a contour image is extracted based on the preprocessed surface image of the battery module.

[0063] The answer is obtained using the following formula:

[0064] ;

[0065] In the formula: For neighborhood size parameters, ; Normalized Gaussian kernel; Fill the boundary with pixels;

[0066] in, , Let represent the Gaussian kernel normalization factor, and (u, v) represent the relative coordinates of the neighborhood centered at (x, y), where u and v are both within the range {−k,−k+1,…,k}. This represents the Gaussian kernel standard deviation. ∈[0.5,2.0];

[0067] in, , This represents the actual width and height of the image. When p < 0, =0, when p≥W =W−1; when q<0, =0, when q≥H, =H−1;

[0068] The above formula utilizes the smoothing properties of the Gaussian kernel to reduce the impact of image noise. The impact of the calculation results is further influenced by the flexibly adjustable... It adapts to battery module images of varying resolutions, resolving the issue of inaccurate calculations caused by a lack of neighborhood information for image boundary pixels, thus enabling... It can more comprehensively and accurately reflect the characteristics of pixels and surrounding areas, providing reliable intermediate data support for preprocessing formulas;

[0069] The storage module is used to store standard battery module image data and receive contour images output by the camera module, and temporarily store the contour images.

[0070] The matching module is used to obtain the contour images stored in the storage module and the temporarily stored contour images, match the two sets of contour images, and then forward the matched contour images to the analysis module to perform analysis operations.

[0071] The matching module has an internal recognition unit, which is used to recognize the marking information of all contour images acquired by the matching module, so that contour images with the same marking information can be matched with each other. That is, each set of matched contour images contains two contour images, which are respectively from the standard battery module image data set stored in the storage module and the temporarily stored contour image set.

[0072] The analysis module receives the matched contour images, evaluates the similarity of the contour images, and analyzes whether the appearance of the battery module is up to standard.

[0073] During the analysis module's operation phase, when analyzing whether the battery module's appearance is up to standard, the comprehensive similarity evaluation result of each matched contour image is used as the quantitative standard for determining whether the analysis is up to standard.

[0074] ;

[0075] In the formula: For comprehensive similarity; Number of groups to match for the contour image; Let be the similarity of the r-th group of contour images; Configure weights for the r-th group of contour images; These are the curvature values ​​at the positions on contour a and contour b, respectively, at the positions on contour b and contour b, at the positions on contour b and contour b, respectively, at the positions on contour a .... This means taking the shortest contour line between the two contours for integration calculation; The bounded area of ​​the intersection region of the two contours; The bounded area of ​​the region formed by the union of the two contours; Let be the perimeter of the intersection region of the two contours; Let be the perimeter of the outer contour of the region where the two contours meet; ∮ is the curve integral along the contour.

[0076] The above formula uses the number of contour image matching groups, and The core of this system is a comprehensive similarity calculation logic that simultaneously considers the shape curvature, spatial area, and boundary perimeter features of the contour, comprehensively reflecting the differences between the standard and the contour of the battery module to be tested; weights are configured. The introduction of this method can adjust the weights according to the importance of different contour areas in the appearance inspection of battery modules, making the overall similarity more in line with actual inspection needs and providing a precise quantitative basis for judging whether the appearance of battery modules is qualified.

[0077] in, During computation, the center points of the two matched contour images are aligned before the computation operation is performed. After obtaining the result, it is compared with the preset qualified threshold in the analysis module. If the battery module's appearance is not less than the preset acceptable threshold, it is considered acceptable; otherwise, it is considered unacceptable.

[0078] The value follows:

[0079] Calculate the mean foreground-to-background ratio of each matched contour image in its source image, denoted as . ,but , express The summation result;

[0080] The above logic determines the ratio of foreground to background in the source image for each set of matched contour images by calculating the mean, and then combining the sum of all the means. The design directly correlates the foreground and background proportions of the contour image with the weights, giving higher weights to contour areas that have a higher proportion in the appearance of the battery module and a greater impact on appearance compliance (such as the main body contour of the battery), and lower weights to areas that have a lower proportion and a smaller impact (such as the contours of small connecting parts). This avoids the problem that traditional fixed weight allocation cannot reflect the differences in importance of different contour areas, making the contribution of each contour area more reasonable when calculating the comprehensive similarity, and further improving the accuracy and reliability of the appearance inspection results.

[0081] The output and interaction module is used to receive the analysis results from the analysis module, sort the currently tested battery modules based on the analysis results, and output the analysis results to the control panel of the battery module production equipment.

[0082] The camera module is connected to a storage module via a wireless network. The storage module is connected to a matching module and an analysis module via a wireless network. The matching module contains an identification unit connected to a wireless network. The analysis module contains an output and interaction module connected to a wireless network.

[0083] In this embodiment, the camera module acquires surface images of the battery module, preprocesses the surface images to extract the contour images corresponding to the surface images of the battery module, the storage module simultaneously stores standard battery module image data and receives the contour images output by the camera module, and temporarily stores the contour images, the matching module runs in the back to acquire the contour images stored in the storage module and the temporarily stored contour images, matches the two sets of contour image sets, and then forwards the matched contour images to the analysis module to perform analysis operations, the recognition unit simultaneously recognizes the marking information of all contour images acquired by the matching module, and makes the contour images with the same marking information match each other, and then the analysis module receives the matched contour images, analyzes whether the appearance of the battery module is qualified based on the contour image similarity evaluation, and finally receives the analysis results of the analysis module through the output and interaction module, sorts the currently inspected battery modules based on the analysis results and outputs the analysis results to the control panel of the battery module production equipment;

[0084] It should be noted that the sorting operation of battery modules in the output and interaction module is performed by a robotic arm module pre-installed next to the conveyor belt assembly of the battery module production equipment. After receiving the analysis results, the robotic arm module sorts the battery modules.

[0085] In the above embodiments, the system can accurately collect and efficiently process images of the battery module surface during actual implementation. It obtains standard and image data to be tested through a standardized process, and then accurately determines whether the appearance of the battery module is qualified through precise matching and comprehensive similarity evaluation. It can also automatically sort and feed back to the production equipment based on the results, which not only improves the detection accuracy and efficiency and reduces human error, but also optimizes the production process in a timely manner, ensures the production quality of battery modules, reduces the risk of defective products flowing out, and helps the production process to be more stable and efficient.

[0086] The following is an application example of the system described in the above embodiments:

[0087] A new energy battery manufacturer has introduced a machine vision-based new energy battery module appearance inspection system to inspect the appearance quality of its square lithium-ion battery modules, ensuring that the products leaving the factory are free from appearance defects such as surface scratches, bulges, and corner deformation.

[0088] After system startup, the initial calibration phase begins. Workers place a standard battery module on the conveyor belt assembly of the production equipment. The conveyor belt transports the standard battery module directly beneath the slip ring, ensuring that the center point of the horizontal plane of the slip ring is on the same vertical line as the center point of the top surface of the standard battery module. The system drives an electric shaft to rotate the slip ring to a preset 30° tilt angle. Then, it controls an electric rail on the sliding rod to push a slider carrying the slip ring downwards until the slip ring surrounds the standard battery module, with the highest point of the slip ring surface not exceeding the height of the top surface of the standard battery module and the lowest point not falling below the surface of the conveyor belt. Subsequently, a high-speed, high-definition camera on the electric rail on the inner wall of the slip ring captures images around the standard battery module at a shooting rate of 30 frames per second and a circular motion speed of 5 cm / s. Simultaneously, a built-in position sensor records the shooting position synchronously at a frequency of 30 times per second. Each captured image of the standard battery module surface is marked with corresponding position information. The camera module preprocesses the captured images to obtain preprocessed images, and finally extracts the contour image set of the standard battery module, storing it in the storage module as standard image data.

[0089] In the actual testing process, the production equipment conveyor belt sequentially transports the battery modules to be tested to the area directly below the slip ring, with the orientation of the battery modules relative to the conveyor belt consistent with that of the standard battery modules. The system performs the same image capture, preprocessing, and contour extraction operations on each module as it does on the standard battery modules, resulting in a set of contour images of the module to be tested. This set of images is also marked with position information and temporarily stored in the storage module.

[0090] The matching module activates the recognition unit to identify the positional marker information of all contour images. It then matches the standard contour images with consistent markers with the contour images to be detected in pairs. Each set of matched images comes from the standard image dataset and the temporary image dataset to be detected, forming a total of 20 sets of matched contour image pairs. These matched images are then forwarded to the analysis module.

[0091] The analysis module first aligns the center points of the two matching contour images in each group, then calculates the similarity of each group of images. The similarity of the 20 groups of images ranges from 0.82 to 0.98. Next, the analysis module determines the configuration weight for each group of matched images, with the weights ranging from 0.04 to 0.06. Subsequently, the overall similarity is calculated to be 0.91. The preset pass threshold in the analysis module is 0.85. Since 0.91 is not less than 0.85, the appearance of the battery module under test is deemed acceptable.

[0092] Finally, the output and interaction module receives the pass / fail judgment result, controls the sorting mechanism to transport the module to the qualified product area, and transmits the test results (including a comprehensive similarity of 0.91 and pass / fail judgment) to the control panel of the battery module production equipment in real time for staff to view; at the same time, the system clears the temporarily stored outline image set of the module to be tested, preparing for the testing of the next battery module to be tested.

[0093] In summary, during operation, the system in the above embodiments acquires surface images of battery modules through specific means and optimizes preprocessing to efficiently extract contour images, ensuring the integrity and accuracy of image information. Furthermore, through a consistent acquisition, positioning, and marking mechanism, it enables more accurate matching of contour images of different battery modules, reducing matching errors. Based on multi-dimensional parameters, it calculates comprehensive similarity and scientifically determines whether the appearance is qualified by combining preset thresholds, improving the reliability of detection results. Temporary image data is promptly cleared after detection to avoid redundant data consuming resources. Simultaneously, the operating parameters can be flexibly adjusted according to detection accuracy requirements to adapt to different detection scenarios, effectively improving the efficiency and quality of battery module appearance inspection, assisting production equipment in achieving precise sorting, and ensuring stable battery module production quality.

[0094] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A machine vision-based appearance inspection system for new energy battery modules, characterized in that, include: The camera module is used to acquire surface images of the battery module and preprocess the surface images of the battery module to extract the contour images corresponding to the surface images of the battery module. The storage module is used to store standard battery module image data and receive contour images output by the camera module, and temporarily store the contour images. The matching module is used to obtain the contour images stored in the storage module and the temporarily stored contour images, match the two sets of contour images, and then forward the matched contour images to the analysis module to perform analysis operations. The analysis module receives the matched contour images, evaluates the similarity of the contour images, and analyzes whether the appearance of the battery module is up to standard. The output and interaction module is used to receive the analysis results from the analysis module, sort the currently tested battery modules based on the analysis results, and output the analysis results to the control panel of the battery module production equipment. The preprocessing of the surface image of the battery module includes: ； In the formula: This represents the pixel value at position (x, y) in the preprocessed image. As a key intermediate variable; This represents the pixel value at position (x, y) in the original image; This is the numerical stability constant, with a value of 1; Specifically, each pixel in the battery module surface image is processed based on the above formula to output a preprocessed battery module surface image, and a contour image is extracted based on the preprocessed battery module surface image; The The answer is obtained using the following formula: ； In the formula: For neighborhood size parameters, ; Normalized Gaussian kernel; Fill the boundary with pixels; in, , Let (u, v) represent the Gaussian kernel normalization factor, and (u, v) represent the relative coordinates of the neighborhood centered at (x, y). Both u and v are within the range of (x, y). Within the range, This represents the Gaussian kernel standard deviation. ∈[0.5,2.0]; in, , This represents the actual width and height of the image. When p < 0, =0, when p≥W, =W−1; when q<0, =0, when q≥H, =H−1; During the operation of the analysis module, when analyzing whether the appearance of the battery module is qualified, the comprehensive similarity evaluation result of each matched contour image is used as the quantitative standard for whether the analysis is qualified: ； In the formula: For comprehensive similarity; Number of groups to match for the contour image; Let be the similarity of the r-th group of contour images; Configure weights for the r-th group of contour images; These are the curvature values ​​at the positions on contour a and contour b, respectively, at the positions on contour b and contour b, at the positions on contour b and contour b, respectively, at the positions on contour a .... This means taking the shortest contour line between the two contours for integration calculation; The bounded area of ​​the intersection region of the two contours; The bounded area of ​​the region formed by the union of the two contours; Let be the perimeter of the intersection region of the two contours; Let be the perimeter of the outer contour of the region where the two contours meet; ∮ is the curve integral along the contour. in, During computation, the center points of the two matched contour images are aligned before the computation operation is performed. After obtaining the result, it is compared with the preset qualified threshold in the analysis module. If the battery module's appearance is not less than the preset acceptable threshold, it is considered acceptable; otherwise, it is considered unacceptable.

2. The new energy battery module appearance inspection system based on machine vision according to claim 1, characterized in that, The camera module includes a slider, a slide ring, and a high-speed high-definition camera; The battery module is output from its production equipment and is conveyed by the conveyor belt assembly of the production equipment to the area directly below the slip ring, so that the center point of the plane of the slip ring in a horizontal state is on the same vertical line as the center point of the top surface of the battery module directly below it. The slide bar is fixedly connected to the ground and located on one side of the production equipment conveyor belt assembly. A slider is installed on the side of the slide bar near the conveyor belt assembly via an electric slide rail. The end of the slider away from the slide bar is connected to a slip ring via an electric rotating shaft. A high-speed high-definition camera is installed on the inner ring wall of the slip ring via an electric slide rail. During the initial operation phase of the system, the slider takes the upper limit position of the electric slide rail on the slide bar as the initial position. When the slip ring is carried by the slider to the initial position, the slip ring is in a horizontal posture. In this process, the battery modules output by the generating device are in the same orientation relative to the surface of the conveyor belt assembly each time.

3. The machine vision-based new energy battery module appearance inspection system according to claim 2, characterized in that, During the initial operation phase of the system, a standard battery module is placed on the conveyor belt assembly of the production equipment and transported by the conveyor belt assembly to the area directly below the slip ring. The rotation angle is set to drive the electric shaft to rotate the slip ring to the target angle. The electric slide rail on the drive slide rod pushes the slider to carry the slip ring downward until the slip ring surrounds the standard battery module. Due to the rotation into an inclined posture, the highest point of the slip ring surface is not higher than the top surface of the battery module, and the lowest point is not lower than the surface of the conveyor belt assembly of the production equipment. A high-speed, high-definition camera moves in a circular motion at a predetermined speed in an electric slide rail on the inner wall of the slip ring, and continuously captures images of the battery module surface at a predetermined shooting rate. The captured images of the battery module surface are simultaneously preprocessed to extract contour images. The resulting set of contour images is the standard battery module image data stored in the storage module. The battery modules to be tested output by the conveyor belt assembly of the production equipment perform the same operation as above to obtain their corresponding contour image set, and send them to the storage module at the same time. The images are temporarily stored in the storage module, and the temporarily stored contour image set is cleared after the system obtains the result of whether it is qualified.

4. The machine vision-based new energy battery module appearance inspection system according to claim 3, characterized in that, When the high-speed high-definition camera in the camera module acquires surface images of different battery modules, the initial acquisition position of each battery module is consistent with the slip ring where the high-speed high-definition camera is located during the acquisition operation of different battery modules. The high-speed high-definition camera has a built-in position sensor, and the operating frequency of the position sensor is consistent with the shooting rate of the high-speed high-definition camera. After each battery module surface image is captured, its positioning result at the same time is used for marking. After the contour image is extracted from the battery module surface image, the contour image uses the marking of its source battery module surface image. The standard battery modules transported by the conveyor belt assembly of the production equipment perform the same operations as above during the acquisition of the contour image set. Among them, the operating frequency of the sensor, the shooting rate of the high-speed high-definition camera, and the predetermined speed of the high-speed high-definition camera in the electric slide rail on the inner wall of the slip ring follow the following rules: the higher the detection accuracy requirement, the faster the operating frequency and shooting rate, and the slower the predetermined speed; conversely, the slower the operating frequency and shooting rate, the faster the predetermined speed.

5. The machine vision-based new energy battery module appearance inspection system according to claim 1, characterized in that, The matching module is equipped with a recognition unit, which is used to recognize the marking information of all contour images acquired by the matching module, so that contour images with consistent marking information can be matched with each other. That is, each set of matched contour images contains two contour images, which are respectively from the standard battery module image data set stored in the storage module and the temporarily stored contour image set.

6. The machine vision-based new energy battery module appearance inspection system according to claim 1, characterized in that, The The value follows: Calculate the mean foreground-to-background ratio of each matched contour image in its source image, denoted as . ,but , express The summation result.

7. The new energy battery module appearance inspection system based on machine vision according to claim 1, characterized in that, The camera module is interconnected with a storage module via a wireless network. The storage module is interconnected with a matching module and an analysis module via a wireless network. The matching module is interconnected with an identification unit via a wireless network. The analysis module is interconnected with an output and interaction module via a wireless network.

Images