A battery defect detection method and system based on multimodal sensor

Abstract

The present invention discloses a battery defect detection method and system based on a multimodal sensor. The method comprises: acquiring an image of a battery to be inspected to obtain an image to be inspected; analyzing the image to be inspected according to a preset image processing algorithm to determine a target area in the image suspected of having a defect and its visual features; scanning the target area to extract tactile features; utilizing a random forest algorithm to obtain a preliminary visual defect probability corresponding to the visual features, and a preliminary tactile defect probability corresponding to the tactile features; and obtaining a target defect probability under target probability distribution according to the D-S theory based on the preliminary visual defect probability and the preliminary tactile defect probability, so as to determine whether a target defect exists in the target area based on the target defect probability. By combining a tactile sensor with image acquisition, the method can multimodally fuse 2D visual data and 3D tactile data, improve the reliability of the final determination, and effectively control the cost of the detection equipment.

Description

Technical Field

[0001] The present invention relates to the technical field of battery detection, and in particular to a battery defect detection method and system based on a multimodal sensor. Background Art

[0002] In battery manufacturing and quality inspection, the integrity of the battery's appearance and the quality of its welds are crucial to the battery's service life, reliability, and overall safety of the end product. During the actual production process, batteries may exhibit a variety of defects, such as scratches, dents, oxidation, and dimensional deviations. These defects can directly lead to serious safety incidents such as battery leakage, fire, and even explosion.

[0003] Currently, existing inspection technologies have many limitations when it comes to inspecting battery appearance and welding quality. While traditional 2D visual inspection technology can detect battery appearance and welding defects to a certain extent, since battery casings are usually made of metal and have highly reflective properties, traditional 2D visual systems are easily interfered with by the reflective effects of the metal surface during the inspection process, resulting in undesirable phenomena such as light spots and shadows in the image, which affects the accuracy and reliability of the inspection and makes it difficult to effectively identify subtle defects on highly reflective surfaces such as aluminum casings. To overcome these shortcomings, some companies are considering using 3D cameras for inspection, which can improve the accuracy and reliability of inspection to a certain extent. However, the high cost of 3D camera equipment undoubtedly adds a huge production cost burden to most battery manufacturers, limiting their large-scale application.

[0004] Therefore, those skilled in the art are in urgent need of developing a new technical solution to solve the above problems. Summary of the Invention

[0005] In order to overcome the problems existing in the related art, the present invention discloses a battery defect detection method and system based on a multimodal sensor.

[0006] According to a first aspect of the disclosed embodiments of the present invention, a battery defect detection method based on a multimodal sensor is provided. The method is applied to a battery defect detection system based on a multimodal sensor. The system includes: an image acquisition module, a tactile sensor module, a processor, and a motion control platform. The method includes:

[0007] Capturing images of the battery to be inspected on the motion control platform by the image acquisition module to obtain images to be inspected;

[0008] Analyzing the image to be inspected by the processor according to a preset image processing algorithm, determining a target area in the image suspected of having defects and extracting visual features corresponding to the target area;

[0009] Scanning the target area by the tactile sensor, and extracting tactile features according to the scanning result of the target area by the processor;

[0010] Obtaining, by the processor, a preliminary visual defect probability corresponding to the visual feature and a preliminary tactile defect probability corresponding to the tactile feature using a random forest algorithm;

[0011] The processor obtains the target defect probability under the target probability distribution according to the visual preliminary defect probability and the tactile preliminary defect probability based on the DS theory, so as to determine whether the target area has a target defect according to the target defect probability.

[0012] Optionally, the processor obtaining, according to the visual preliminary defect probability and the tactile preliminary defect probability, a target defect probability under target probability distribution based on DS theory, so as to determine whether the target area has a target defect according to the target defect probability, includes:

[0013] Converting the visual preliminary defect probability and the tactile preliminary defect probability into the visual defect basic probability distribution and the tactile defect basic probability distribution of the DS theory, and calculating the conflict coefficient between the image acquisition module and the tactile sensor;

[0014] If the conflict coefficient is greater than or equal to a preset coefficient threshold, the basic probability distribution of visual defects and the basic probability distribution of tactile defects are modified by the weight coefficient of the image acquisition module and the weight coefficient of the tactile sensor, and the target defect probability under the target probability distribution is obtained according to the modified basic probability distribution of visual defects and the modified basic probability distribution of tactile defects;

[0015] If the conflict coefficient is less than a preset coefficient threshold, obtaining the target defect probability under the target probability distribution according to the basic probability distribution of visual defects and the basic probability distribution of tactile defects;

[0016] If the target defect probability is greater than or equal to a preset defect threshold corresponding to the target defect, it is determined that the target defect exists in the target area;

[0017] If the target defect probability is less than a preset defect threshold corresponding to the target defect, it is determined that the target defect does not exist in the target area.

[0018] Optionally, the obtaining the target defect probability under the target probability distribution according to the corrected basic probability distribution of visual defects and the corrected basic probability distribution of tactile defects, or the obtaining the target defect probability under the target probability distribution according to the basic probability distribution of visual defects and the basic probability distribution of tactile defects, includes:

[0019] According to the corrected basic probability distribution of visual defects and the corrected basic probability distribution of tactile defects, or according to the basic probability distribution of visual defects and the basic probability distribution of tactile defects, the target defect probability is determined by a preset target defect probability formula,

[0020] Among them, the target defect probability formula is:

[0021] , is the target defect probability in the target probability distribution, is the probability of a defect existing in the basic probability distribution of visual defects, is the probability of the defect existing in the basic probability of the tactile defect, and K is the conflict coefficient.

[0022] Optionally, the correcting the basic probability distribution of visual defects and the basic probability distribution of tactile defects by using the image acquisition module weight coefficient and the tactile sensor weight coefficient includes:

[0023] The basic probability distribution of visual defects is modified by the weight coefficient of the image acquisition module to obtain the modified basic probability distribution of visual defects ,in, is the probability of a defect existing in the basic probability distribution of visual defects, is the weight coefficient of the image acquisition module;

[0024] The basic probability distribution of tactile defects is modified by the tactile sensor weight coefficient to obtain a modified basic probability distribution of tactile defects ,in, is the probability of a defect in the basic probability of tactile defects, is the tactile sensor weight coefficient.

[0025] Optionally, the motion control platform includes a robotic arm and a placement platform, the placement platform is used to place the battery to be tested, and the robotic arm is used to move the battery to be tested; and the image acquisition module is used to capture an image of the battery to be tested on the motion control platform to obtain the image to be tested, including:

[0026] The battery to be inspected on the placement table is captured by an image acquisition module to obtain an image of a detection surface of the battery to be inspected;

[0027] The battery to be inspected is moved by the robot arm during the image acquisition process to ensure that the images to be inspected of all inspection surfaces of the battery to be inspected are acquired.

[0028] Optionally, a plane coordinate system is provided on the placement table, and scanning the target area by the tactile sensor includes:

[0029] Determining the coordinate position of the target area in the plane coordinate system;

[0030] determining a contact path of the tactile sensor when scanning the target area according to the coordinate position;

[0031] The target area is scanned by the tactile sensor according to the contact path.

[0032] According to a second aspect of the disclosed embodiments of the present invention, a battery defect detection system based on a multimodal sensor is provided. The system is used to implement the battery defect detection method based on a multimodal sensor described in the first aspect of the disclosed embodiments of the present invention. The system includes: an image acquisition module, a tactile sensor module, a processor, and a motion control platform;

[0033] The image acquisition module and the tactile sensor module are arranged above the motion control platform and are electrically connected to the processor respectively. The motion control platform is used to place the battery to be detected.

[0034] Optionally, the motion control platform includes a robotic arm and a placement platform, the placement platform is used to place the battery to be tested, and the robotic arm is used to move the battery to be tested.

[0035] Optionally, the robotic arm is used to move the battery to be inspected during the image acquisition process to ensure that the images of each inspection surface of the battery to be inspected are acquired; and after the processor analyzes the image of the battery to be inspected through an image processing algorithm and determines the target area, the battery to be inspected is moved to align the inspection surface corresponding to the target area with the tactile sensor.

[0036] Optionally, a plane coordinate system is provided on the placement table.

[0037] In summary, the technical solutions in the disclosed embodiments of the present invention can bring the following beneficial effects:

[0038] (1) First, use random forest to classify visual features and tactile features, output preliminary defect probability, and then use DS theory to fuse the confidence of multiple sensors and adopt multimodal fusion algorithm to improve the reliability of the final judgment;

[0039] (2) Using low-priced devices such as high-resolution industrial cameras and array tactile sensors as image acquisition modules and tactile sensors, compared with expensive 3D cameras, effectively controls the cost of battery defect detection;

[0040] (3) The “suspected defects” in the target area are further verified by using the tactile sensor, and the area that the tactile sensor focuses on is concentrated in the target area. This can solve the problem of low working efficiency of the tactile sensor and effectively improve the efficiency of battery defect detection.

[0041] Other features and advantages disclosed in the present invention will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings are used to provide a further understanding of the present disclosure and constitute a part of the specification. Together with the following detailed description, they are used to explain the present disclosure but do not constitute a limitation of the present disclosure. In the accompanying drawings:

[0043] Figure 1 is a flow chart showing a battery defect detection method based on a multimodal sensor according to an exemplary embodiment;

[0044] Figure 2 is based on Figure 1 A schematic flow chart of an image acquisition method is shown;

[0045] Figure 3 is based on Figure 1 A schematic flow chart of a method for scanning a target area using a tactile sensor is shown;

[0046] Figure 4 is based on Figure 1 A schematic flow chart of a target defect determination method is shown;

[0047] Figure 5 The figure is a schematic structural diagram of a battery defect detection system based on a multimodal sensor according to an exemplary embodiment. DETAILED DESCRIPTION

[0048] The following is a detailed description of the specific embodiments disclosed in the present invention in conjunction with the accompanying drawings. It should be understood that the specific embodiments described herein are only used to illustrate and explain the present invention and are not intended to limit the present invention.

[0049] The battery defect detection method disclosed in the embodiment of the present invention collects multi-dimensional battery casing data through a combination of visual inspection and tactile inspection. The multi-dimensional data (2D visual data + 3D tactile data) is then fused to obtain high-precision detection results and identify defects on the battery casing surface. The battery defect detection method is applied to a battery defect detection system based on a multimodal sensor. The system includes: an image acquisition module, a tactile sensor module, a processor, and a motion control platform. The image acquisition module and the tactile sensor module are respectively used to collect data in the visual dimension and the tactile dimension. The processor is used to analyze the above data and obtain defect detection results after fusing the data. The motion control platform can be used to place the battery casing to be inspected and move the battery casing so that each surface to be inspected is aligned with the image acquisition module or the tactile sensor.

[0050] Figure 1 FIG. 1 is a flow chart showing a battery defect detection method based on a multimodal sensor according to an exemplary embodiment. Figure 1 As shown, the method includes:

[0051] In step 101, the image acquisition module is used to acquire an image of the battery to be inspected on the motion control platform to obtain an image to be inspected.

[0052] For example, the motion control platform includes a robotic arm and a placement platform. The placement platform is used to place the battery to be tested, and the robotic arm is used to move the battery to be tested. It is understood that the housing of each battery to be tested includes multiple test surfaces. When the image acquisition module captures the image to be tested, it is necessary to fully capture the image of each test surface.

[0053] Specifically, Figure 2 is based on Figure 1 A flow chart of an image acquisition method is shown in FIG. Figure 2 As shown, step 101 includes:

[0054] In step 1011, the battery to be inspected on the placement table is captured by an image acquisition module to obtain an image of a detection surface of the battery to be inspected.

[0055] In step 1012, the battery to be inspected is moved by the robot arm during the image acquisition process to ensure that the images to be inspected of all inspection surfaces of the battery to be inspected are acquired.

[0056] For example, the battery to be inspected on the placement table is moved by a robotic arm so that each inspection surface is aligned with the image acquisition module, thereby completing image acquisition of the inspection surface.

[0057] In step 102, the processor analyzes the image to be inspected according to a preset image processing algorithm to determine a target area in the image where a defect is suspected to exist and extract visual features corresponding to the target area.

[0058] For example, when analyzing the image to be inspected, the processor analyzes each inspection surface separately. Preset image processing algorithms include existing image processing methods such as multi-light source imaging algorithms, edge enhancement algorithms, and contour extraction algorithms. The specific steps for analyzing the image to be inspected according to the preset image processing algorithms are as follows: for each inspection surface, images of the inspection surface captured by the image acquisition module at different angles are fused to highlight defect features. The image gradient of the inspection surface image is calculated using the Sobel or Scharr operator to enhance linear features (for example, scratches typically appear as long, high-gradient regions). Contour extraction techniques are used to determine the contours of the weld area, and local contrast enhancement and other methods are employed to highlight recessed areas. Furthermore, oxidation areas, dimensional deviation areas, weld porosity areas, crack areas, weld slag areas, and cold weld areas on the battery casing surface can also be determined using existing image processing algorithms, and will not be further described in the disclosed embodiments of the present invention. It should be understood that the aforementioned weld areas, recessed areas, weld porosity areas, and crack areas are all target areas suspected of containing defects. The reason why the above target areas are called suspected defects is that the 2D visual data collected by the image acquisition module is easily affected by reflections from the metal surface and other ambient light, resulting in certain deviations in the detection results. To ensure the accuracy of the defect detection results on each inspection surface of the battery casing, the defect detection results obtained through 2D visual data are called suspected defect target areas, and the "suspected defects" are further verified through the following steps 103-105.

[0059] The visual features corresponding to the target area include: texture features (LBP, Haralick), shape descriptors (Hu moments), and deep learning features (CNN intermediate layer output).

[0060] In step 103, the target area is scanned by the tactile sensor, and the tactile features are extracted by the processor according to the scanning result of the target area.

[0061] For example, tactile sensors can be used to further verify suspected defects in the target area, identifying contact force / pressure distribution in the weld zone, cold welds, structural deformation, scratches, pits, and deformed bumps. Tactile features corresponding to the target area include: pressure distribution mean / variance, force time series characteristics (zero-crossing rate), and surface roughness.

[0062] The placement table is provided with a plane coordinate system, specifically, Figure 3 is based on Figure 1 A flowchart diagram of a method for scanning a target area using a tactile sensor is shown in FIG. Figure 3 As shown, step 103 includes:

[0063] In step 1031 , the coordinate position of the target area in the plane coordinate system is determined.

[0064] In step 1032 , a contact path of the tactile sensor when scanning the target area is determined according to the coordinate position.

[0065] In step 1033 , the target area is scanned according to the contact path by the tactile sensor.

[0066] For example, the plane coordinate system uses the lower left corner of the placement platform as its center, the lower edge of the placement platform as the x-axis, and the left edge of the placement platform as the y-axis. Each point on the placement platform can be assigned a corresponding coordinate position in the plane coordinate system. After determining the coordinate position corresponding to the target area, a reasonable tactile sensor scanning contact path can be set based on the location and size of the target area, ensuring that the tactile sensor scans along the specified contact path.

[0067] In addition, it should be noted that after determining the target area, the robot arm can move the battery to be tested to align the detection surface corresponding to the target area with the tactile sensor, so that the tactile sensor can perform the next scanning action.

[0068] In step 104 , the processor uses a random forest algorithm to obtain a preliminary visual defect probability corresponding to the visual feature, and obtains a preliminary tactile defect probability corresponding to the tactile feature.

[0069] For example, a random forest classifier is established, and the visual features and tactile features are used as inputs of the random forest classifier respectively, and the visual preliminary defect probability and the tactile preliminary defect probability are obtained according to the output of the random forest classifier.

[0070] In step 105, the processor obtains the target defect probability under the target probability distribution according to the visual preliminary defect probability and the tactile preliminary defect probability based on the DS theory, so as to determine whether there is a target defect in the target area according to the target defect probability.

[0071] For example, according to the preliminary visual defect probability and the preliminary tactile defect probability, they are converted into output confidence under the DS theory to obtain the probability distribution of the target defect, thereby determining whether the target defect exists in the target area. Among them, the full name of the DS theory is the Dempster-Shafer theory. It is a method for dealing with uncertainty problems, which describes uncertainty information through "interval estimation" rather than "point estimation". In the disclosed embodiment of the present invention, random forests are first used to classify visual features and tactile features, and preliminary defect probabilities are output. Then, the DS theory is used to fuse the confidence of multiple sensors, and a multimodal fusion algorithm is used to improve the reliability of the final judgment.

[0072] Specifically, Figure 4 is based on Figure 1 A flow chart of a target defect determination method is shown in FIG. Figure 4 As shown, step 105 includes:

[0073] In step 1051, the preliminary visual defect probability and the preliminary tactile defect probability are converted into the basic visual defect probability distribution and the basic tactile defect probability distribution of the DS theory, and the conflict coefficient between the image acquisition module and the tactile sensor is calculated.

[0074] For example, the calculation formula of the conflict coefficient K is: , The probability of the defect existing in the basic probability distribution of the visual defect, is the probability of a defect within the basic probability of the tactile defect.

[0075] In step 1052, if the conflict coefficient is greater than or equal to the preset coefficient threshold, the basic probability distribution of visual defects and the basic probability distribution of tactile defects are corrected by the image acquisition module weight coefficient and the tactile sensor weight coefficient, and the target defect probability under the target probability distribution is obtained according to the corrected basic probability distribution of visual defects and the corrected basic probability distribution of tactile defects.

[0076] For example, if the conflict coefficient is greater than or equal to the preset coefficient threshold, it means that there is a high degree of contradiction between the 2D visual data collected by the image acquisition module and the 3D tactile data collected by the tactile sensor. At this time, the basic probability distribution of visual defects and the basic probability distribution of tactile defects need to be corrected.

[0077] Specifically, the basic probability distribution of visual defects and the basic probability distribution of tactile defects are modified by the weight coefficient of the image acquisition module and the weight coefficient of the tactile sensor, including:

[0078] The basic probability distribution of visual defects is corrected by the weight coefficient of the image acquisition module to obtain the corrected basic probability distribution of visual defects ,in, The probability of the defect existing in the basic probability distribution of the visual defect, is the weight coefficient of the image acquisition module.

[0079] The basic probability distribution of tactile defects is corrected by the tactile sensor weight coefficient to obtain the corrected basic probability distribution of tactile defects ,in, is the probability of the defect existing in the basic probability of the tactile defect, is the tactile sensor weight coefficient. The image acquisition module weight coefficient and the tactile sensor weight coefficient are obtained through experiments.

[0080] In step 1053, if the conflict coefficient is less than the preset coefficient threshold, the target defect probability under the target probability distribution is obtained according to the basic probability distribution of visual defects and the basic probability distribution of tactile defects.

[0081] For example, if the conflict coefficient is less than the preset coefficient threshold, it means that there is no high degree of contradiction between the 2D visual data collected by the image acquisition module and the 3D tactile data collected by the tactile sensor, and there is no need to correct the basic probability distribution of visual defects and the basic probability distribution of tactile defects. The target defect probability can be directly determined by the basic probability distribution of visual defects and the basic probability distribution of tactile defects.

[0082] Specifically, obtaining the target defect probability under the target probability distribution according to the corrected basic probability distribution of visual defects and the corrected basic probability distribution of tactile defects, or obtaining the target defect probability under the target probability distribution according to the basic probability distribution of visual defects and the basic probability distribution of tactile defects, includes:

[0083] According to the corrected basic probability distribution of visual defects and the corrected basic probability distribution of tactile defects, or according to the basic probability distribution of visual defects and the basic probability distribution of tactile defects, the target defect probability is determined by a preset target defect probability formula,

[0084] Among them, the target defect probability formula is:

[0085] , is the target defect probability in the target probability distribution, The probability of the defect existing in the basic probability distribution of the visual defect, is the probability of the defect existing in the basic probability of the tactile defect, and K is the conflict coefficient.

[0086] In step 1054 , if the target defect probability is greater than or equal to the preset defect threshold corresponding to the target defect, it is determined that the target defect exists in the target area.

[0087] In step 1055 , if the target defect probability is less than the preset defect threshold corresponding to the target defect, it is determined that the target defect does not exist in the target area.

[0088] For example, it can be understood that the types of target defects include: welds, dents, oxidation, welding slag, cold welds, etc. The annotation library contains preset defect thresholds corresponding to each target type. By analyzing 2D visual data + 3D tactile data, the target defect type can be preliminarily determined. If the target defect probability is greater than or equal to the preset defect threshold corresponding to the target defect type, it is determined that the target defect exists in the target area. Otherwise, it is determined that the target defect does not exist in the target area. Based on the preset defect thresholds in the standard library, a high-quality annotation library (covering the main defect types) can be established in the early stage. Active learning can be introduced in the medium term to reduce incremental costs. In the long term, efficient and sustainable model optimization can be achieved by combining data and transfer learning.

[0089] Figure 5 FIG. 1 is a structural diagram of a battery defect detection system based on a multimodal sensor according to an exemplary embodiment. Figure 5 As shown, the system 500 includes: an image acquisition module 510, a tactile sensor module 520, a processor (not shown in the figure) and a motion control platform (including a robotic arm 531 and a placement table 532); the image acquisition module 510 and the tactile sensor module 520 are arranged above the motion control platform (an additional image acquisition module 510 can be added to the side of the motion control platform to capture images of other inspection surfaces of the battery to be inspected), and are respectively electrically connected to the processor. The motion control platform is used to place the battery to be inspected.

[0090] Specifically, the image acquisition module 510 can be a high-resolution industrial camera with a circularly polarized light source to suppress reflections and structured light. It can also emit light of different colors and angles to suppress reflections. The tactile sensor 520 can be an array tactile sensor. High-resolution industrial cameras and array tactile sensors are relatively affordable, effectively controlling costs during battery defect detection.

[0091] The image acquisition module 510 is used to capture images of the battery to be inspected on the motion control platform and obtain the image to be inspected; the tactile sensor 520 is used to scan the target area; the processor is used to analyze the image to be inspected according to a preset image processing algorithm, determine the target area in the image that is suspected of having defects and extract the visual features corresponding to the target area; extract tactile features based on the scanning results of the target area; use the random forest algorithm to obtain the visual preliminary defect probability corresponding to the visual features, and obtain the tactile preliminary defect probability corresponding to the tactile features; based on the visual preliminary defect probability and the tactile preliminary defect probability, obtain the target defect probability under the target probability distribution according to the DS theory, so as to determine whether the target area has a target defect based on the target defect probability.

[0092] The motion control platform includes a robotic arm 531 and a placement table 532. The placement table 532 is used to place the battery to be tested, and the robotic arm 531 is used to move the battery to be tested. The battery to be tested is moved during the image acquisition process to ensure that the images of each detection surface of the battery to be tested are acquired. After the processor analyzes the image of the battery to be tested through an image processing algorithm and determines the target area, the battery to be tested is moved to align the detection surface corresponding to the target area with the tactile sensor.

[0093] A plane coordinate system is provided on the placement table. A coordinate position of the target area in the plane coordinate system is determined; a contact path of the tactile sensor when scanning the target area is determined based on the coordinate position; and the target area is scanned by the tactile sensor based on the contact path.

[0094] In summary, the present invention discloses a battery defect detection method and system based on a multimodal sensor, which includes: performing image acquisition on the battery to be inspected to obtain the image to be inspected; analyzing the image to be inspected according to a preset image processing algorithm to determine the target area and visual features suspected of having defects in the image; scanning the target area to extract tactile features; using a random forest algorithm to obtain the visual preliminary defect probability corresponding to the visual features, and obtaining the tactile preliminary defect probability corresponding to the tactile features; based on the visual preliminary defect probability and the tactile preliminary defect probability, obtaining the target defect probability under the target probability distribution according to the DS theory, so as to determine whether the target area has a target defect based on the target defect probability. By combining a tactile sensor with image acquisition, the multimodal fusion of 2D visual data and 3D tactile data can be achieved, thereby improving the reliability of the final judgment and effectively controlling the cost of the detection equipment.

[0095] The preferred embodiments of the present disclosure are described in detail above in conjunction with the accompanying drawings. However, the present disclosure is not limited to the specific details of the above embodiments. Within the technical concept of the present disclosure, various simple modifications can be made to the technical solutions of the present disclosure, and these simple modifications all fall within the scope of protection of the present disclosure.

[0096] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any appropriate manner without contradiction. In order to avoid unnecessary repetition, the present disclosure will not further describe various possible combinations.

[0097] In addition, the various embodiments of the present disclosure may be arbitrarily combined, and as long as they do not violate the concept of the present disclosure, they should also be regarded as the contents disclosed by the present disclosure.

Claims

1. A battery defect detection method based on a multimodal sensor, characterized in that: The method is applied to a battery defect detection system based on a multimodal sensor, the system comprising: an image acquisition module, a tactile sensor module, a processor, and a motion control platform, the method comprising: Capturing images of the battery to be inspected on the motion control platform by the image acquisition module to obtain images to be inspected; Analyzing the image to be inspected by the processor according to a preset image processing algorithm, determining a target area in the image suspected of having defects and extracting visual features corresponding to the target area; Scanning the target area by the tactile sensor, and extracting tactile features according to the scanning result of the target area by the processor; Obtaining, by the processor, a preliminary visual defect probability corresponding to the visual feature and a preliminary tactile defect probability corresponding to the tactile feature using a random forest algorithm; Converting the visual preliminary defect probability and the tactile preliminary defect probability into the visual defect basic probability distribution and the tactile defect basic probability distribution of the DS theory, and calculating the conflict coefficient between the image acquisition module and the tactile sensor; If the conflict coefficient is greater than or equal to a preset coefficient threshold, the basic probability distribution of visual defects and the basic probability distribution of tactile defects are modified by the weight coefficient of the image acquisition module and the weight coefficient of the tactile sensor, and the target defect probability under the target probability distribution is obtained according to the modified basic probability distribution of visual defects and the modified basic probability distribution of tactile defects; If the conflict coefficient is less than a preset coefficient threshold, obtaining the target defect probability under the target probability distribution according to the basic probability distribution of visual defects and the basic probability distribution of tactile defects; The correcting of the basic probability distribution of visual defects and the basic probability distribution of tactile defects by using the image acquisition module weight coefficient and the tactile sensor weight coefficient includes: The basic probability distribution of visual defects is modified by the weight coefficient of the image acquisition module to obtain the modified basic probability distribution of visual defects ,in, is the probability of a defect existing in the basic probability distribution of visual defects, is the weight coefficient of the image acquisition module; The basic probability distribution of tactile defects is modified by the tactile sensor weight coefficient to obtain a modified basic probability distribution of tactile defects ,in, is the probability of a defect in the basic probability distribution of the tactile defect, is the tactile sensor weight coefficient; The target defect probability is determined by the preset target defect probability formula. Among them, the target defect probability formula is: , is the target defect probability in the target probability distribution, is the probability of a defect existing in the basic probability distribution of visual defects, is the probability of the defect existing in the basic probability of the tactile defect, and K is the conflict coefficient.

2. The battery defect detection method based on a multimodal sensor according to claim 1, characterized in that: The method further comprises: If the target defect probability is greater than or equal to a preset defect threshold corresponding to the target defect, it is determined that the target defect exists in the target area; If the target defect probability is less than a preset defect threshold corresponding to the target defect, it is determined that the target defect does not exist in the target area.

3. The battery defect detection method based on a multimodal sensor according to claim 2, characterized in that: The target defect probability formula is used to: The target defect probability is determined by a preset target defect probability formula based on the corrected basic probability distribution of visual defects and the corrected basic probability distribution of tactile defects, or the target defect probability is determined by a preset target defect probability formula based on the basic probability distribution of visual defects and the basic probability distribution of tactile defects.

4. The battery defect detection method based on a multimodal sensor according to claim 1, characterized in that: The motion control platform includes a robotic arm and a placement platform, the placement platform is used to place the battery to be tested, and the robotic arm is used to move the battery to be tested; the image acquisition module is used to capture an image of the battery to be tested on the motion control platform to obtain the image to be tested, including: The battery to be inspected on the placement table is captured by an image acquisition module to obtain an image of a detection surface of the battery to be inspected; The battery to be inspected is moved by the robot arm during the image acquisition process to ensure that the images to be inspected of all inspection surfaces of the battery to be inspected are acquired.

5. The battery defect detection method based on a multimodal sensor according to claim 4, characterized in that: A plane coordinate system is provided on the placement table, and scanning the target area by the tactile sensor includes: Determining the coordinate position of the target area in the plane coordinate system; determining a contact path of the tactile sensor when scanning the target area according to the coordinate position; The target area is scanned by the tactile sensor according to the contact path.

6. A battery defect detection system based on a multimodal sensor, characterized in that: The system is used to implement the battery defect detection method based on a multimodal sensor according to any one of claims 1 to 5, and the system comprises: an image acquisition module, a tactile sensor module, a processor, and a motion control platform; The image acquisition module and the tactile sensor module are arranged above the motion control platform and are electrically connected to the processor respectively. The motion control platform is used to place the battery to be detected.

7. The battery defect detection system based on multimodal sensors according to claim 6, characterized in that: The motion control platform includes a mechanical arm and a placement platform, wherein the placement platform is used to place the battery to be tested, and the mechanical arm is used to move the battery to be tested.

8. The battery defect detection system based on multimodal sensors according to claim 7, characterized in that: The robotic arm is used to move the battery to be inspected during the image acquisition process to ensure that the images of each inspection surface of the battery to be inspected are acquired, and after the processor analyzes the image of the battery to be inspected through an image processing algorithm and determines the target area, the robotic arm is used to move the battery to be inspected to align the inspection surface corresponding to the target area with the tactile sensor.

9. The battery defect detection system based on multimodal sensors according to claim 7, characterized in that: A plane coordinate system is provided on the placement platform.

Images