Lithium battery pack polarity intelligent detection system and method based on deep learning

Abstract

The invention relates to the technical field of new energy battery production detection, in particular to a lithium battery pack polarity intelligent detection system and method based on deep learning, and the system comprises a hardware execution module which comprises an industrial personal computer, an adjustable illumination assembly, a sound-light alarm device, a high-precision industrial camera with a gigabit Ethernet interface, and an adjustable support; the software algorithm module comprises an upper computer interaction interface, a deep learning battery cell detection algorithm based on a YOLO model and a multi-target matching analysis algorithm; according to the lithium battery pack polarity intelligent detection system and method based on deep learning, the problems that traditional manual detection is low in efficiency and high in omission ratio, can only depend on a fixed template and cannot adapt to battery core model difference and battery pack difference are effectively solved.

Description

Technical Field

[0001] The present invention relates to the technical field of new energy battery production detection technology, and specifically to a deep learning-based intelligent detection system and method for lithium battery pack polarity. Background Art

[0002] With the explosive growth of the new energy battery market, power battery pack safety has become a core concern in the industry. Battery packs typically consist of dozens to hundreds of cells connected in series and parallel. Improper polarity alignment can directly lead to serious safety hazards such as short circuits and thermal runaway. Traditional battery pack polarity detection systems rely on manual or simple mechanical devices, resulting in slow detection and prone to errors. Therefore, in response to this situation, there is an urgent need to develop a deep learning-based intelligent lithium battery pack polarity detection system and method to overcome the shortcomings of current practical applications. Summary of the Invention

[0003] The purpose of the present invention is to provide a lithium battery pack polarity intelligent detection system and method based on deep learning to solve the problems raised in the above background technology.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A deep learning-based intelligent detection system for lithium battery pack polarity, comprising:

[0006] Hardware execution module: industrial computer, adjustable lighting components, sound and light alarm device, high-precision industrial camera with Gigabit Ethernet interface and adjustable bracket;

[0007] Software algorithm module: host computer interactive interface, deep learning cell detection algorithm based on YOLO model, and multi-target matching analysis algorithm;

[0008] The industrial computer is directly connected to the industrial camera via Gigabit Ethernet and is connected to the sound and light alarm device via the RS232 serial port;

[0009] The industrial camera is used to collect battery pack images and transmit them to the industrial computer;

[0010] The deep learning battery cell detection algorithm is used to identify the polarity position of the battery cell;

[0011] The multi-objective matching analysis algorithm is used to match the real-time image with the standard polarity arrangement in the sample library;

[0012] The host computer interactive interface is used to display the matching results and the position of the wrong battery cell, and trigger the sound and light alarm.

[0013] As a further solution of the present invention: the industrial camera is the MV-CS200-10GM model, which supports automatic / manual adjustment of gain, exposure time, white balance, and integrates ISP algorithm and CCM / LSC correction function.

[0014] As a further solution of the present invention: the industrial computer calls the camera API through the SDK to send a photo-taking instruction, and sends an ASCII instruction through the serial port to control the light and buzzer of the sound and light alarm device.

[0015] As a further solution of the present invention: the adjustable bracket supports height adjustment to adapt to different production scenarios.

[0016] As a further solution of the present invention: the implementation steps of the deep learning battery cell detection algorithm include:

[0017] Collect battery cell images and perform data annotation;

[0018] Train polarity detection model based on YOLO algorithm;

[0019] Call the model to output the polarity position and direction of the battery cell.

[0020] As a further solution of the present invention: the multi-target matching analysis algorithm calculates the position deviation and outputs the matching result by comparing the real-time image with the standard polar arrangement coordinates of the sample library.

[0021] As a further solution of the present invention: the host computer interaction interface includes: a camera connection control module, a sample library acquisition module and a real-time image detection result display module.

[0022] An intelligent detection method for a lithium battery pack polarity intelligent detection system based on deep learning as described above comprises the following steps:

[0023] S1, collect battery pack images through industrial cameras and transmit them to industrial computers via Gigabit Ethernet;

[0024] S2. Using a deep learning battery cell detection algorithm, the polarity position of each battery cell in the image is identified based on the YOLO model.

[0025] S3. Using a multi-target matching analysis algorithm, the polarity arrangement of the battery cells in the real-time image is matched with the standard polarity arrangement stored in the sample library;

[0026] S4. If the matching fails, the wrong cell position will be displayed on the host computer interface and an ASCII command will be sent through the RS232 serial port to trigger an audible and visual alarm.

[0027] As a further solution of the present invention: in step S2, the implementation of the deep learning battery cell detection algorithm includes:

[0028] Collect battery cell images and perform data annotation;

[0029] Train polarity detection model based on YOLO algorithm;

[0030] Call the model to output the polarity position and direction of the battery cell.

[0031] As a further solution of the present invention: in step S4, triggering the sound and light alarm includes: the industrial computer sends ASCII instructions through the RS232 serial port to control the light and buzzer of the sound and light alarm device.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] 1. The present invention uses an industrial computer, an adjustable lighting assembly, an audible and visual alarm device, a high-precision industrial camera with a Gigabit Ethernet interface, and an adjustable bracket to design a device for detecting the polarity arrangement of lithium batteries in a variety of cell types and battery packs. This effectively solves the problems of traditional manual detection, such as low efficiency, high missed detection rate, reliance on fixed templates, and inability to adapt to differences in cell types and battery packs.

[0034] 2. The YOLO-based visual inspection technology of the present invention can quickly and accurately detect the polarity of battery packs, greatly improving production efficiency. Through the automated inspection system, manual intervention can be reduced, reducing labor costs. At the same time, the defective rate caused by detection errors can be reduced, thereby reducing material waste and production costs. Accurate polarity detection is the key to ensuring battery pack performance and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a block diagram of a lithium battery pack polarity intelligent detection system based on deep learning in an embodiment of the present invention.

[0036] Figure 2 This is a block diagram of the cell detection algorithm and target matching algorithm in an embodiment of the present invention.

[0037] Figure 3 This is a block diagram of the connection between an industrial computer and an industrial camera in an embodiment of the present invention.

[0038] Figure 4 This is a block diagram of the connection between the industrial computer and the buzzer alarm in an embodiment of the present invention. DETAILED DESCRIPTION

[0039] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0040] The specific implementation of the present invention is described in detail below with reference to specific embodiments.

[0041] See also Figure 1-Figure 4 , an embodiment of the present invention provides a lithium battery pack polarity intelligent detection system based on deep learning, comprising:

[0042] Hardware execution module: industrial computer, adjustable lighting components, sound and light alarm device, high-precision industrial camera with Gigabit Ethernet interface and adjustable bracket;

[0043] Software algorithm module: host computer interactive interface, deep learning cell detection algorithm based on YOLO model, and multi-target matching analysis algorithm;

[0044] The industrial computer is directly connected to the industrial camera via Gigabit Ethernet and is connected to the sound and light alarm device via the RS232 serial port;

[0045] The industrial camera is used to collect battery pack images and transmit them to the industrial computer;

[0046] The deep learning battery cell detection algorithm is used to identify the polarity position of the battery cell;

[0047] The multi-objective matching analysis algorithm is used to match the real-time image with the standard polarity arrangement in the sample library;

[0048] The host computer interactive interface is used to display the matching results and the position of the wrong battery cell, and trigger the sound and light alarm.

[0049] In one embodiment of the present invention, see Figure 1-Figure 4 The industrial camera is the MV-CS200-10GM model, which supports automatic / manual adjustment of gain, exposure time, and white balance, and integrates ISP algorithm and CCM / LSC correction function.

[0050] The industrial computer calls the camera API through the SDK to send a photo-taking instruction, and sends an ASCII instruction through the serial port to control the light and buzzer of the sound and light alarm device.

[0051] The adjustable bracket supports height adjustment to adapt to different production scenarios. The adjustable bracket is an existing technology, for example, a telescopic cylinder is used to adjust the height of the bracket, etc., which will not be described in detail here.

[0052] The implementation steps of the deep learning battery cell detection algorithm include:

[0053] Collect battery cell images and perform data annotation;

[0054] Train polarity detection model based on YOLO algorithm;

[0055] Call the model to output the polarity position and direction of the battery cell.

[0056] The multi-target matching analysis algorithm compares the real-time image with the standard polar arrangement coordinates of the sample library, calculates the position deviation and outputs the matching result.

[0057] The host computer interaction interface includes: a camera connection control module, a sample library acquisition module and a real-time image detection result display module.

[0058] The present invention provides an intelligent detection method for a lithium battery pack polarity intelligent detection system based on deep learning as described above, comprising the following steps:

[0059] S1: Capture battery pack images through an industrial camera and transmit them to an industrial computer via Gigabit Ethernet;

[0060] S2: Using a deep learning battery cell detection algorithm, the polarity position of each battery cell in the image is identified based on the YOLO model;

[0061] S3: Using a multi-target matching analysis algorithm, the polarity arrangement of the battery cells in the real-time image is matched with the standard polarity arrangement stored in the sample library;

[0062] S4: If the matching fails, the wrong cell position will be displayed on the host computer interface and an ASCII command will be sent through the RS232 serial port to trigger an audible and visual alarm.

[0063] In one embodiment of the present invention, see Figure 1-Figure 4 In step S2, the implementation of the deep learning battery cell detection algorithm includes:

[0064] Collect battery cell images and perform data annotation;

[0065] Train polarity detection model based on YOLO algorithm;

[0066] Call the model to output the polarity position and direction of the battery cell.

[0067] In step S4, triggering the sound and light alarm includes: the industrial computer sends ASCII instructions through the RS232 serial port to control the light and buzzer of the sound and light alarm device.

[0068] Specifically, the present invention's deep learning-based intelligent polarity detection system for lithium-ion battery packs consists of two modules: a hardware execution mechanism and a software algorithm platform. The hardware execution module includes an industrial computer, adjustable lighting components, an audio and visual alarm device, a high-precision industrial camera, and an adjustable bracket. The software algorithm module integrates a host computer interface, a deep learning detection algorithm, and a multi-target matching analysis algorithm. Through multi-dimensional collaborative operation, the system can achieve high-speed and accurate detection of lithium-ion battery pack polarity. The details are as follows:

[0069] The industrial camera model is the MV-CS200-10GM, which uses the IMX183 CMOS, has a 20-megapixel resolution, and supports Gigabit Ethernet (GigE Vision V2.0 protocol, compatible with the GenICam standard). It supports automatic or manual adjustment of gain, exposure time, white balance, LUT, and gamma correction, integrates multiple ISP algorithms, supports CCM and LSC correction, and is rich in features.

[0070] The light source model is OPT-FLT600800-HW, and the adjustable bracket can adjust the height to adapt to more production scenarios;

[0071] The host computer's function is to connect to the camera for real-time image acquisition, obtain samples, and save the correctly arranged samples into the sample library. It then checks the polarity arrangement of subsequent batteries. If it passes, it will display a successful match. If not, it will display a failed match and indicate the location of the incorrect battery cell.

[0072] The battery cell detection algorithm mainly collects and annotates data, trains a model based on the Yolo algorithm, and then writes the battery cell detection algorithm based on the model; the target matching algorithm matches the real-time image with the collected samples;

[0073] The training data source for the Yolo algorithm training model is: 1,000 samples of different battery cells and battery packs were collected;

[0074] Labeling rules: positive pole is brown, negative pole is brown;

[0075] Parameter settings: input image resolution 5472×3648, number of iterations: 300 epochs.

[0076] The industrial computer is directly connected to the LAN interface of the industrial camera via the built-in Gigabit Ethernet port; the industrial computer is connected to the buzzer alarm via RS232 communication; the industrial computer calls the camera API through the SDK, sends a trigger photo command and receives image data; the industrial computer sends ASCII commands through the serial port to control the alarm light and buzzer; the software part of the host computer interface includes camera connection, sample collection and image detection, etc.

[0077] Among them, the alarm instruction format is: the alarm turns on the green light [0xa0, 0x02, 0x01, 0xa3], the alarm turns off the green light [0xa0, 0x02, 0x00, 0xa2], the alarm turns on the red light and buzzes [0xa0, 0x07, 0x02, 0xa9], the alarm turns off the red light and buzzer.

[0078] In summary, the present invention adopts an industrial computer, an adjustable lighting component, an audible and visual alarm device, a high-precision industrial camera with a gigabit Ethernet interface, and an adjustable bracket, etc., to design a device for detecting the polarity arrangement of lithium batteries that can detect a variety of battery cells and a variety of battery packs. This effectively solves the problems of low efficiency, high missed detection rate, reliance on fixed templates, and inability to adapt to differences in battery cell models and battery packs in traditional manual detection. The visual inspection technology based on YOLO in the present invention can quickly and accurately detect the polarity of battery packs, greatly improving production efficiency. Through the automated detection system, manual intervention can be reduced and labor costs can be reduced. At the same time, the defective rate caused by detection errors can be reduced, thereby reducing material waste and production costs. Accurate polarity detection is the key to ensuring battery pack performance and safety.

[0079] The actual operation is as follows:

[0080] Test object: 18650-5S6P battery pack (30 cells); Test samples: 500 sets (including 50 defective samples)

[0081] The results are compared in the following table:

[0082]

[0083]

[0084] It should be noted that, in the present invention, it should be understood that although this specification is described in terms of implementation methods, not every implementation method contains only one independent technical solution. This narrative method of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

Claims

1. A lithium battery pack polarity intelligent detection system based on deep learning, characterized in that: include: Hardware execution module: industrial computer, adjustable lighting components, sound and light alarm device, high-precision industrial camera with Gigabit Ethernet interface and adjustable bracket; Software algorithm module: host computer interactive interface, deep learning cell detection algorithm based on YOLO model, and multi-target matching analysis algorithm; The industrial computer is directly connected to the industrial camera via Gigabit Ethernet and is connected to the sound and light alarm device via the RS232 serial port; The industrial camera is used to collect battery pack images and transmit them to the industrial computer; The deep learning battery cell detection algorithm is used to identify the polarity position of the battery cell; The multi-objective matching analysis algorithm is used to match the real-time image with the standard polarity arrangement in the sample library; The host computer interactive interface is used to display the matching results and the position of the wrong battery cell, and trigger the sound and light alarm.

2. The deep learning-based intelligent detection system for lithium battery polarity according to claim 1, characterized in that: The industrial camera is the MV-CS200-10GM model, which supports automatic / manual adjustment of gain, exposure time, and white balance, and integrates ISP algorithm and CCM / LSC correction functions.

3. The lithium battery pack polarity intelligent detection system based on deep learning according to claim 1, characterized in that: The industrial computer calls the camera API through the SDK to send a photo-taking instruction, and sends an ASCII instruction through the serial port to control the light and buzzer of the sound and light alarm device.

4. The lithium battery pack polarity intelligent detection system based on deep learning according to claim 1, characterized in that: The adjustable bracket supports height adjustment to adapt to different production scenarios.

5. The lithium battery pack polarity intelligent detection system based on deep learning according to claim 1, characterized in that: The implementation steps of the deep learning battery cell detection algorithm include: Collect battery cell images and perform data annotation; Train polarity detection model based on YOLO algorithm; Call the model to output the polarity position and direction of the battery cell.

6. The lithium battery pack polarity intelligent detection system based on deep learning according to claim 1, characterized in that: The multi-target matching analysis algorithm compares the real-time image with the standard polar arrangement coordinates of the sample library, calculates the position deviation and outputs the matching result.

7. The lithium battery pack polarity intelligent detection system based on deep learning according to claim 1, characterized in that: The host computer interaction interface includes: a camera connection control module, a sample library acquisition module and a real-time image detection result display module.

8. An intelligent detection method for a lithium battery polarity intelligent detection system based on deep learning according to any one of claims 1 to 7, characterized in that: The following steps are involved: S1, collect battery pack images through industrial cameras and transmit them to industrial computers via Gigabit Ethernet; S2. Using a deep learning battery cell detection algorithm, the polarity position of each battery cell in the image is identified based on the YOLO model. S3. Using a multi-target matching analysis algorithm, the polarity arrangement of the battery cells in the real-time image is matched with the standard polarity arrangement stored in the sample library; S4. If the matching fails, the wrong cell position will be displayed on the host computer interface and an ASCII command will be sent through the RS232 serial port to trigger an audible and visual alarm.

9. The intelligent detection method of the lithium battery pack polarity intelligent detection system based on deep learning according to claim 8, characterized in that: In step S2, the implementation of the deep learning battery cell detection algorithm includes: Collect battery cell images and perform data annotation; Train polarity detection model based on YOLO algorithm; Call the model to output the polarity position and direction of the battery cell.

10. The intelligent detection method of the lithium battery polarity intelligent detection system based on deep learning according to claim 8, characterized in that: In step S4, triggering the sound and light alarm includes: the industrial computer sends ASCII instructions through the RS232 serial port to control the light and buzzer of the sound and light alarm device.

Images