Visual identification method and device for battery connecting piece

By acquiring and preprocessing the RGB diagram and depth diagram, and calculating the target depth value and real coordinate value of the battery connector, the problem of insufficient deep data accuracy of RGBD equipment is solved, and the efficiency and safety of disassembly of retired power batteries is improved.

CN118941759APending Publication Date: 2024-11-12WUHAN UNIV OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411006905.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-25
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

In the prior art, the deep data obtained through RGBD equipment is not very accurate and has a lot of noise, resulting in low efficiency and safety of disassembly of retired power batteries.

Method used

Provide a visual recognition method for battery connectors, by obtaining RGB diagrams and multiple depth diagrams of battery connectors, pre-processing to obtain high-quality depth images, and combining RGB diagrams and pixel coordinates, the target depth value of battery connectors is calculated. and the true coordinate value, thereby improving the accuracy and efficiency of disassembly.

Benefits of technology

By improving the accuracy of depth images and removing noise, higher battery connector recognition and position accuracy are achieved, and the efficiency and safety of removable power batteries are improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118941759A_ABST
    Figure CN118941759A_ABST
Patent Text Reader

Abstract

The invention relates to a battery connecting piece visual identification method and device, and belongs to the technical field of decommissioning power battery disassembly, the method comprises the following steps: obtaining an RGB image and a plurality of depth maps of a battery connecting piece, thus preprocessing deep data of the plurality of depth maps, and obtaining a high-quality depth image with improved precision and noise removal; furthermore, the pixel coordinate of the central point of the battery connecting piece can be obtained according to the RGB image, so that the target depth value of the battery connecting piece can be obtained according to the high-quality depth image, the RGB image and the pixel coordinate, and then the real coordinate value of the battery connecting piece can be obtained according to the pixel coordinate and the target depth value. Therefore, the battery connecting piece can be disassembled according to the real coordinate value, and the efficiency of disassembling the retired power battery is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of disassembly of retired power batteries, and in particular to a method and device for visually identifying battery connectors. Background Art

[0002] The disassembly of power batteries is the first key step in battery recycling. The efficiency and safety of disassembly will directly affect the recycling benefits of batteries. Due to the complexity and diversity of the disassembly work itself, the accuracy requirements for visual identification and positioning technology are constantly increasing during the disassembly process. Considering the insufficient accuracy of the two-dimensional images and deep data obtained by RGBD camera equipment, the technology combining 2D and 3D visual methods becomes crucial. The 2D image provides details and texture information, combined with the spatial position data provided by 3D deep data, which can significantly improve the accuracy and stability of identification and positioning. This fusion of 2D and 3D visual methods can not only effectively cope with complex disassembly tasks, but also provide higher reliability in a changing working environment. However, the deep data obtained by some RGBD equipment has the problems of low accuracy and high noise, which has a great impact on the subsequent deep data processing.

[0003] Therefore, it is urgent to propose a battery connector visual recognition method and device to solve the technical problems in the prior art that the deep data obtained by RGBD equipment is of low accuracy and high noise, resulting in low efficiency and safety in the disassembly of retired power batteries. Summary of the invention

[0004] In view of this, it is necessary to provide a battery connector visual recognition method and device to solve the technical problems in the prior art that the deep data obtained by RGBD equipment is of low accuracy and high noise, resulting in low efficiency and safety in the disassembly of retired power batteries.

[0005] In order to solve the above problems, the present invention provides a battery connector visual recognition method, comprising: Acquire an RGB image and multiple depth images of the battery connector, and preprocess the multiple depth images to obtain a high-quality depth image; According to the RGB image, the pixel coordinates of the center point of the battery connector are obtained; Obtaining a target depth value of the battery connector according to the high-quality depth image, the RGB image and the pixel coordinates; The real coordinate value of the battery connector is obtained according to the pixel coordinates and the target depth value.

[0006] In a possible implementation, obtaining the pixel coordinates of the center point of the battery connector according to the RGB image includes: Detecting the RGB image to obtain a region of interest of the battery connector; The region of interest is processed to obtain the pixel coordinates of the center point of the battery connector.

[0007] In a possible implementation, obtaining a target depth value of the battery connector according to the high-quality depth image, the RGB image, and the pixel coordinates includes: Performing depth filling on the high-quality depth image to obtain filled depth data; Aligning the filled-in deep layer data with the RGB image to obtain an accurately matched depth image; A target depth value of the battery connector is obtained according to the pixel coordinates and the precise matching depth map.

[0008] In a possible implementation manner, preprocessing the multiple depth maps to obtain a high-quality depth image includes: Obtaining an initial depth map according to an average depth of each pixel on the multiple depth maps; Calculating according to the initial depth map and the initial depth value of each pixel in the multiple depth maps to obtain a weight of each pixel; Adjusting the weight according to the initial depth value to obtain the target weight of each pixel; A high-quality depth image is obtained according to the initial depth value of each pixel point in the multiple depth images and all target weights.

[0009] In a possible implementation manner, adjusting the weight according to the initial depth value to obtain the target weight of each pixel includes: Taking a weighted average of the initial depth value and the weight of each pixel in the multiple depth maps to obtain a target depth map; Obtaining a local variance evaluation value according to the pixel values ​​of all pixels in the target depth map; The weight is adjusted according to the local variance evaluation value to obtain the target weight of each pixel.

[0010] In a possible implementation, obtaining the initial depth map according to the average depth of each pixel on the multiple depth maps includes: Determine, according to the multiple depth maps, an initial depth value of each pixel point in each depth map; Obtaining an average depth value according to all initial depth values ​​of each pixel; The initial depth map is obtained based on the average value of all depths of all pixels.

[0011] In a possible implementation, the processing the region of interest to obtain the pixel coordinates of the center point of the battery connector includes: Performing edge detection on the region of interest to obtain edge information of the battery connector; The edge information is searched based on a contour detection algorithm to obtain the pixel coordinates of the center point of the battery connector.

[0012] In a possible implementation, obtaining a target depth value of the battery connector according to the pixel coordinates and the precise matching depth map includes: Mapping the pixel coordinates to the precise matching depth map to obtain a target pixel point; A target depth value of the battery connector is obtained according to the target pixel point.

[0013] In a possible implementation, the calculation of each pixel in the high-quality depth image is:

[0014] In the formula, Represents the pixel value of each pixel in the high-quality depth image; Indicates i Pixels in the depth map p The target weight of Indicates i Pixels in the depth map p The initial depth value of .

[0015] On the other hand, the present invention also provides a battery connector visual identification device, comprising: An image acquisition module, used to acquire an RGB image and multiple depth images of the battery connector, and preprocess the multiple depth images to obtain a high-quality depth image; A center point weight module, used to obtain the pixel coordinates of the center point of the battery connector according to the RGB image; a depth value calculation module, configured to obtain a target depth value of the battery connector according to the high-quality depth image, the RGB image and the pixel coordinates; A coordinate value determination module is used to obtain the real coordinate value of the battery connector according to the pixel coordinate and the target depth value.

[0016] The beneficial effect of the present invention is to obtain an RGB image and multiple depth maps of a battery connector, so that the deep data of the multiple depth maps can be preprocessed to obtain a high-quality depth image with improved accuracy and noise removal; further, the pixel coordinates of the center point of the battery connector can be obtained according to the RGB image, so that the target depth value of the battery connector can be obtained according to the high-quality depth image, the RGB image and the pixel coordinates, and then the real coordinate value of the battery connector can be obtained according to the pixel coordinates and the target depth value, so that the battery connector can be disassembled according to the real coordinate value, thereby improving the efficiency of disassembly of retired power batteries. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic flow chart of an embodiment of a battery connector visual identification method provided by the present invention; Figure 2 For the present invention Figure 1 A schematic flow chart of an embodiment of step S101; Figure 3 For the present invention Figure 2 A schematic flow chart of an embodiment of step S201; Figure 4 For the present invention Figure 2 A schematic flow chart of an embodiment of step S203; Figure 5 For the present invention Figure 1 A schematic flow chart of an embodiment of step S103; Figure 6 A schematic structural diagram of an embodiment of a battery connector visual identification device provided by the present invention; Figure 7 A schematic structural diagram of an embodiment of an electronic device provided by the present invention. DETAILED DESCRIPTION

[0018] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0019] An RGBD device is a camera device that simultaneously acquires color images and depth information. It combines the functions of an RGB camera and a depth sensor. This device can provide richer, more three-dimensional visual data and has a miniaturized design, making it easy to integrate into mobile devices or other systems.

[0020] Battery connectors refer to conductors used to carry current between components in a battery circuit, such as connectors between single cells, or between battery terminals and battery pack terminals.

[0021] like Figure 1 As shown, a specific embodiment of the present invention discloses a battery connector visual recognition method, comprising: S101, obtaining an RGB image and multiple depth images of a battery connector, and preprocessing the multiple depth images to obtain a high-quality depth image; S102, obtaining the pixel coordinates of the center point of the battery connector according to the RGB image; S103, obtaining a target depth value of the battery connector according to the high-quality depth image, the RGB image and the pixel coordinates; S104 , obtaining the real coordinate value of the battery connector according to the pixel coordinate and the target depth value.

[0022] It should be understood that: it can be applied to retired power batteries to automatically disassemble retired power batteries of automobiles. The RGB image can be obtained by an RGB camera, and multiple depth images can be detected by a depth sensor of an RGBD device.

[0023] In a specific embodiment of the present invention, the camera shooting surface is ensured to be horizontal, the battery connector is ensured to be within the shooting range, the camera is controlled, and the image acquisition operation is performed. In the process of collecting deep data, it is found that the position of the noise generated in each frame is different. Therefore, the camera device can be kept stationary during sampling, and N frames of depth maps of the same size can be continuously collected. Then, the N frames of depth maps of the same size can be preprocessed to obtain a high-quality depth image. The pixel coordinates of the center point of the battery connector can also be obtained based on the RGB image, and then the target depth value of the battery connector in the high-quality depth image can be obtained by using the high-quality depth image, RGB image and pixel coordinates; then the real coordinate value of the battery connector can be obtained based on the pixel coordinates and the target depth value.

[0024] Compared with the prior art, the present embodiment provides an RGB image and multiple depth images of a battery connector, so that the deep data of the multiple depth images can be preprocessed to obtain a high-quality depth image with improved accuracy and noise removed; further, the pixel coordinates of the center point of the battery connector can be obtained according to the RGB image, so that the target depth value of the battery connector can be obtained according to the high-quality depth image, the RGB image and the pixel coordinates, and then the real coordinate value of the battery connector can be obtained according to the pixel coordinates and the target depth value, so that the battery connector can be disassembled according to the real coordinate value, thereby improving the efficiency of disassembly of retired power batteries.

[0025] In some embodiments of the present invention, Figure 2 As shown, step S101 includes: S201, obtaining an initial depth map according to the average depth of each pixel point on multiple depth maps; S202, calculating according to the initial depth map and the initial depth value of each pixel in the multiple depth maps to obtain the weight of each pixel; S203, adjusting the weight according to the initial depth value to obtain a target weight for each pixel; S204: Obtain a high-quality depth image according to the initial depth value of each pixel in the multiple depth images and all target weights.

[0026] In a specific embodiment of the present invention, in order to achieve a better depth filling effect for the deep map, a new preprocessing algorithm is proposed in the embodiment of the present invention. The algorithm effectively eliminates noise and improves the quality of deep data through adaptive adjustment and feedback mechanism. Specifically, the algorithm performs weighted average fusion on multiple depth maps, uses newly generated deep data to evaluate the noise level, and optimizes the weights again based on the feedback quality evaluation results, and finally generates high-quality deep data. This method comprehensively considers local quality and adaptive adjustment, and is suitable for application scenarios of denoising and improving the quality of deep data.

[0027] In some embodiments of the present invention, Figure 3 , step S201 includes: S301, determining an initial depth value of each pixel in each depth map according to multiple depth maps; S302, obtaining an average depth value according to all initial depth values ​​of each pixel; S303 : Obtain an initial depth map according to the average values ​​of all depths of all pixels.

[0028] In a specific embodiment of the present invention, since the multiple depth maps are depth images of the same content and size, the pixels are also the same. After obtaining the multiple depth maps, the initial depth value of each pixel on each depth map can be determined, and then the initial depth value of each pixel on the multiple depth maps can be calculated. Then, a new initial depth map can be generated based on the average value of all depths of all pixels. , calculated as shown in formula (1): (1) In the formula, Represents the pixel point in the initial depth map p The depth value of Indicates i Pixels in the depth map p The initial depth value of N Indicates the number of depth maps.

[0029] Then, based on the quality evaluation index of the initial depth map, each pixel is weighted to further eliminate the influence of noise. The calculation of is shown in formula (2): (2) In some embodiments of the present invention, Figure 4 , step S203 includes: S401, performing weighted averaging on the initial depth value and weight of each pixel in multiple depth maps to obtain a target depth map; S402, obtaining a local variance evaluation value according to the pixel values ​​of all pixels in the target depth map; S403: Adjust the weight according to the local variance evaluation value to obtain the target weight of each pixel.

[0030] In a specific embodiment of the present invention, each pixel point can be p Perform weighted averaging to generate a fused target depth map , calculated as shown in formula (3): (3) In the formula, Represents the fused target depth map Medium pixel p The depth value of .

[0031] Then, the local variance evaluation value can be obtained according to the pixel values ​​of all pixels in the target depth map, and the calculation is shown in formula (4): (4) In the formula, Represents pixel p The local variance at Represents pixel p fields, Represents the pixel value of the pixel in the area, Represents the mean value of the pixel value in the area, k Indicates the number of pixels in the area.

[0032] Based on the initial fused target depth map The quality evaluation of the target depth map is further adjusted to form an adaptive feedback mechanism. Local quality evaluation index , = .

[0033] The local quality evaluation index can be based on the feedback , adjust the weight of each pixel. The weight adjustment is shown in formula (5): (5) In the formula, and It is an adjustment parameter that can be optimized based on experimental results.

[0034] According to the initial depth value of each pixel in multiple depth maps and all target weights, a high-quality depth image is generated. The calculation of each pixel in the high-quality depth image is shown in formula (6): (6) In the formula, Represents the pixel value of each pixel in the high-quality depth image; Indicates i Pixels in the depth map p The target weight of Indicates i Pixels in the depth map p The initial depth value of .

[0035] In some embodiments of the present invention, step S102 includes: Detect the RGB image to obtain the region of interest of the battery connector; The region of interest is processed to obtain the pixel coordinates of the center point of the battery connector.

[0036] In a specific embodiment of the present invention, the neural network technology can be used to detect the RGB image, obtain the range of the battery connector, and use it as the region of interest (ROI), so as to reduce redundant information in the RGB image; the image processing algorithm can also be used to process the region of interest in the RGB image to obtain the pixel coordinates of the center point of the battery connector.

[0037] In some embodiments of the present invention, processing the region of interest to obtain the pixel coordinates of the center point of the battery connector includes: Perform edge detection on the region of interest to obtain edge information of the battery connector; The edge information is searched based on the contour detection algorithm to obtain the pixel coordinates of the center point of the battery connector.

[0038] In a specific embodiment of the present invention, the Canny algorithm is used to perform edge detection on the ROI area in the RGB image to obtain edge information of the battery connector, and then the contour detection algorithm is used to find the pixel coordinates of the center point of the bolt connector of the battery from the results of the Canny edge detection.

[0039] In some embodiments of the present invention, Figure 5 As shown, step S103 includes: S501, performing depth filling on the high-quality depth image to obtain filled depth data; S502, registering and aligning the filled deep layer data and the RGB image to obtain an accurate matching depth map; S503 : Obtain a target depth value of the battery connector according to the pixel coordinates and the accurate matching depth map.

[0040] In a specific embodiment of the present invention, there may be missing areas in the high-quality depth image due to occlusion or sensor range limitation, and the details still need to be enhanced to improve the overall depth map quality. Therefore, the deep data needs to be further filled and enhanced. The depth filling algorithm can more accurately combine the texture, color and other information in the RGB image to predict and fill the missing areas in the high-quality depth image, so the high-quality depth image can be depth filled by the depth filling algorithm to obtain filled deep data, and then the filled deep data and the RGB image can be aligned to obtain an accurately matched depth map to ensure the accurate match of the filled deep data with the RGB image, thereby providing an accurate data basis for the detection and determination of the position in the subsequent steps.

[0041] In some embodiments of the present invention, step S503 includes: Map the pixel coordinates to the precise matching depth map to obtain the target pixel point; According to the target pixel point, the target depth value of the battery connector is obtained.

[0042] In a specific embodiment of the present invention, after obtaining the pixel coordinates of the battery connector in the RGB image, the pixel coordinates are mapped to the precise matching depth map to obtain the target pixel point, and according to the pixel position of the target pixel point, the depth value of the pixel position is extracted from the precise matching depth map.

[0043] Furthermore, the real coordinate value of the battery connector can be calculated based on the pixel coordinates, the camera intrinsic parameters and the target depth value, wherein the camera intrinsic parameters can be set according to the actual situation of the camera, and then the spatial coordinates of the center point of the battery connector relative to the camera, that is, the real coordinate value, can be calculated, as shown in formula (7): (7) In the formula, u and v are pixel coordinates, X , Y and Z is the three-dimensional space coordinate, T It is the camera internal parameter.

[0044] The embodiments of the present invention can fully utilize the detail information in the RGB image and the spatial information in the deep data by combining 2D image processing with 3D deep information, thereby achieving higher recognition accuracy and position accuracy. This application of 2D and 3D fusion is crucial because it not only makes up for the lack of accuracy of RGBD equipment when used alone, but also provides more stable and accurate recognition and positioning effects in complex environments; further, an innovative preprocessing algorithm is introduced to effectively eliminate noise in the depth image and improve image quality. The embodiments of the present invention not only improve work efficiency and accuracy, but also reduce costs and delays caused by recognition errors, bringing important technological advances and economic benefits to industrial production.

[0045] In order to better implement the battery connector visual recognition method in the embodiment of the present invention, based on the battery connector visual recognition method, the embodiment of the present invention also provides a battery connector visual recognition device, such as Figure 6 As shown, the battery connector visual identification device 600 includes: An image acquisition module 601 is used to acquire an RGB image and multiple depth images of the battery connector, and pre-process the multiple depth images to obtain a high-quality depth image; A center point weight module 602 is used to obtain the pixel coordinates of the center point of the battery connector according to the RGB image; A depth value calculation module 603 is used to obtain a target depth value of the battery connector according to the high-quality depth image, the RGB image and the pixel coordinates; The coordinate value determination module 604 is used to obtain the real coordinate value of the battery connector according to the pixel coordinate and the target depth value.

[0046] The battery connector visual identification device 600 provided in the above embodiment can implement the technical solution described in the above battery connector visual identification method embodiment. The specific implementation principles of the above modules or units can refer to the corresponding contents in the above battery connector visual identification method embodiment, which will not be repeated here.

[0047] like Figure 7 As shown, the present invention also provides an electronic device 700. The electronic device 700 includes a processor 701, a memory 702 and a display 703. Figure 7 Only some components of the electronic device 700 are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0048] In some embodiments, the memory 702 may be an internal storage unit of the electronic device 700, such as a hard disk or memory of the electronic device 700. In other embodiments, the memory 702 may also be an external storage device of the electronic device 700, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc., equipped on the electronic device 700.

[0049] Furthermore, the memory 702 may include both an internal storage unit of the electronic device 700 and an external storage device. The memory 702 is used to store application software installed in the electronic device 700 and various data.

[0050] In some embodiments, the processor 701 may be a central processing unit (CPU), a microprocessor or other data processing chip, and is used to run program codes or process data stored in the memory 702, such as the battery connector visual recognition method of the present invention.

[0051] In some embodiments, the display 703 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 703 is used to display information of the electronic device 700 and to display a visual user interface. The components 701-703 of the electronic device 700 communicate with each other via a system bus.

[0052] In some embodiments of the present invention, when the processor 701 executes the battery connector visual recognition program in the memory 702, the following steps may be implemented: Acquire an RGB image and multiple depth images of the battery connector, and preprocess the multiple depth images to obtain a high-quality depth image; According to the RGB image, the pixel coordinates of the center point of the battery connector are obtained; Obtain a target depth value of the battery connector according to the high-quality depth image, the RGB image, and the pixel coordinates; According to the pixel coordinates and the target depth value, the real coordinate value of the battery connector is obtained.

[0053] It should be understood that: when the processor 701 executes the battery connector visual recognition program in the memory 702, in addition to the above functions, other functions can also be implemented. For details, please refer to the description of the corresponding method embodiment above.

[0054] Furthermore, the embodiment of the present invention does not specifically limit the type of the electronic device 700 mentioned, and the electronic device 700 may be a portable electronic device such as a mobile phone, a tablet computer, a personal digital assistant (PDA), a wearable device, a laptop computer, etc. Exemplary embodiments of portable electronic devices include but are not limited to portable electronic devices equipped with IOS, Android, Microsoft or other operating systems. The above-mentioned portable electronic devices may also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, the electronic device 700 may not be a portable electronic device, but a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0055] Accordingly, an embodiment of the present application also provides a computer-readable storage medium, which is used to store computer-readable programs or instructions. When the program or instructions are executed by a processor, the steps or functions of the battery connector visual identification method provided in the above-mentioned method embodiments can be implemented.

[0056] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware (such as a processor, a controller, etc.) through a computer program, and the computer program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0057] The battery connector visual identification method and device provided by the present invention are introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for visually identifying battery connectors, It is characterized in that include: Acquire an RGB image and multiple depth images of the battery connector, and preprocess the multiple depth images to obtain a high-quality depth image; According to the RGB image, the pixel coordinates of the center point of the battery connector are obtained; Obtaining a target depth value of the battery connector according to the high-quality depth image, the RGB image and the pixel coordinates; The real coordinate value of the battery connector is obtained according to the pixel coordinates and the target depth value.

2. The battery connector visual recognition method according to claim 1, It is characterized in that The step of obtaining the pixel coordinates of the center point of the battery connector according to the RGB image includes: Detecting the RGB image to obtain a region of interest of the battery connector; The region of interest is processed to obtain the pixel coordinates of the center point of the battery connector.

3. The battery connector visual recognition method according to claim 1, It is characterized in that The step of obtaining a target depth value of the battery connector according to the high-quality depth image, the RGB image and the pixel coordinates includes: Performing depth filling on the high-quality depth image to obtain filled depth data; Aligning the filled-in deep layer data with the RGB image to obtain an accurately matched depth image; A target depth value of the battery connector is obtained according to the pixel coordinates and the precise matching depth map.

4. The battery connector visual recognition method according to claim 1, It is characterized in that The preprocessing of the multiple depth maps to obtain a high-quality depth image includes: Obtaining an initial depth map according to an average depth of each pixel on the multiple depth maps; Calculating according to the initial depth map and the initial depth value of each pixel in the multiple depth maps to obtain a weight of each pixel; Adjusting the weight according to the initial depth value to obtain the target weight of each pixel; A high-quality depth image is obtained according to the initial depth value of each pixel point in the multiple depth images and all target weights.

5. The battery connector visual recognition method according to claim 4, It is characterized in that The step of adjusting the weight according to the initial depth value to obtain the target weight of each pixel includes: Taking a weighted average of the initial depth value and the weight of each pixel in the multiple depth maps to obtain a target depth map; Obtaining a local variance evaluation value according to the pixel values ​​of all pixels in the target depth map; The weight is adjusted according to the local variance evaluation value to obtain the target weight of each pixel.

6. The battery connector visual recognition method according to claim 4, It is characterized in that The step of obtaining an initial depth map according to the average depth of each pixel on the multiple depth maps comprises: Determine, according to the multiple depth maps, an initial depth value of each pixel point in each depth map; Obtaining an average depth value according to all initial depth values ​​of each pixel; The initial depth map is obtained based on the average value of all depths of all pixels.

7. The battery connector visual recognition method according to claim 2, It is characterized in that The processing of the region of interest to obtain the pixel coordinates of the center point of the battery connector includes: Performing edge detection on the region of interest to obtain edge information of the battery connector; The edge information is searched based on a contour detection algorithm to obtain the pixel coordinates of the center point of the battery connector.

8. The battery connector visual recognition method according to claim 3, It is characterized in that The step of obtaining a target depth value of the battery connector according to the pixel coordinates and the accurate matching depth map includes: Mapping the pixel coordinates to the precise matching depth map to obtain a target pixel point; A target depth value of the battery connector is obtained according to the target pixel point.

9. The battery connector visual recognition method according to claim 1, It is characterized in that The calculation of each pixel in the high-quality depth image is: In the formula, Represents the pixel value of each pixel in the high-quality depth image; Indicates i Pixels in the depth map p The target weight of Indicates i Pixels in the depth map p The initial depth value of .

10. A battery connector visual recognition device, It is characterized in that include: An image acquisition module, used to acquire an RGB image and multiple depth images of the battery connector, and preprocess the multiple depth images to obtain a high-quality depth image; A center point weight module, used to obtain the pixel coordinates of the center point of the battery connector according to the RGB image; a depth value calculation module, configured to obtain a target depth value of the battery connector according to the high-quality depth image, the RGB image and the pixel coordinates; A coordinate value determination module is used to obtain the real coordinate value of the battery connector according to the pixel coordinate and the target depth value.