Damage detection system and method for lower protective plate of electric vehicle

Abstract

The invention provides an electric vehicle lower protection plate damage detection system and method, and relates to the technical field of electric vehicles. The system comprises at least one camera module, an image processing module and an early warning module. The camera shooting module is installed at a fixed position below a road, and when the electric automobile passes through the position above the camera shooting module, a lower protection plate image of the electric automobile is shot. The image processing module is used for carrying out damage detection on the lower protection plate image based on a trained lower protection plate detection model, and when the detection result is that the lower protection plate of the electric vehicle is damaged, the damage position and the damage degree are sent to the early warning module; and the early warning module sends early warning information to a corresponding vehicle or a vehicle management system, wherein the early warning information comprises vehicle information, a damage position and a damage degree. According to the invention, based on the trained lower protection plate detection model, damage detection can be carried out on the lower protection plate of the electric vehicle in real time, when the lower protection plate is damaged, an alarm is given in time, and spontaneous combustion accidents of the electric vehicle can be prevented.

Description

Technical Field

[0001] The present application relates to the technical field of electric vehicles, and in particular to a system and method for detecting damage to an underbody guard plate of an electric vehicle. Background Art

[0002] With the adjustment of the global energy structure and the enhancement of environmental awareness, electric vehicles are gradually replacing traditional fuel vehicles and becoming the mainstream of future transportation. However, the accompanying safety issues of electric vehicles have also attracted widespread attention, especially the risks related to batteries and their thermal management systems. The battery pack of electric vehicles generates a lot of heat during charging and discharging, and the chassis liquid cooling plate, as its main heat dissipation device, plays a vital role. The integrity of the liquid cooling plate not only affects the performance and life of the battery, but is also directly related to the driving safety of the vehicle.

[0003] When electric vehicles are driving on complex roads, the chassis underbody guard plate is very susceptible to impact, wear and tear from road obstacles, and even damage. Once the underbody guard plate is damaged, it is very easy to affect the liquid cooling plate, causing the battery thermal management system to fail, and the battery temperature will not be effectively controlled. When the battery is charging, overcharging may cause the battery temperature to rise rapidly, increasing the risk of high-temperature spontaneous combustion, and then causing major safety accidents such as vehicle fires. Relevant accident cases show that about 80% of car owners are unaware of the damage to the underbody guard plate, and the liquid cooling plate inside the underbody guard plate only needs to deform 5mm to be judged as a greater risk and needs to be disassembled for further inspection and repair. In recent years, with the rapid development of the electric vehicle market, similar high-temperature spontaneous combustion accidents have occurred frequently, further highlighting the urgency of underbody guard plate detection.

[0004] At present, most electric vehicle underbody panel inspections on the market rely on manual visual inspections or regular maintenance. However, these methods are inefficient, have long inspection cycles, and cannot meet the real-time safety needs of vehicles while driving. Traditional inspection technologies, such as pressure monitoring and thermal imaging, although they perform well in certain scenarios, usually cannot accurately reflect the surface deformation or structural damage of the underbody panel, and there is a risk of missed inspections. Manual inspections are not only time-consuming and labor-intensive, but are also easily interfered with by human factors, making it difficult to ensure the comprehensiveness and accuracy of the inspections. In this context, how to accurately, efficiently, and in real time detect damage to the underbody panel of electric vehicles has become a technical problem that needs to be solved urgently. Summary of the invention

[0005] The purpose of this application is to provide a system and method for detecting damage to the underside guard plate of an electric vehicle, which can improve the accuracy and real-time performance of detecting damage to the underside guard plate of an electric vehicle.

[0006] In a first aspect, an embodiment of the present application provides a system for detecting damage to an underbody guard of an electric vehicle, comprising:

[0007] At least one camera module, an image processing module and an early warning module; the at least one camera module is connected to the image processing module respectively, and the image processing module is connected to the early warning module;

[0008] The camera module is installed at a fixed position below the road. When the electric vehicle passes over the camera module, the image of the lower guard plate of the electric vehicle is captured, and the image of the lower guard plate is sent to the image processing module.

[0009] The image processing module is used to perform damage detection on the lower guard plate image based on the trained lower guard plate detection model, and when the detection result shows that the lower guard plate of the electric vehicle is damaged, the damage location and damage degree are sent to the early warning module;

[0010] The early warning module is used to obtain vehicle information of the electric vehicle. When receiving the detection result that the lower guard plate of the electric vehicle is damaged, it sends early warning information to the corresponding vehicle or vehicle management system. The early warning information includes vehicle information, damage location and damage degree.

[0011] In a possible implementation, before the image processing module performs damage detection on the lower guard plate image based on the trained lower guard plate detection model, it is further used to:

[0012] The lower guard plate image is preprocessed, and the preprocessing includes: grayscale, denoising and contrast enhancement.

[0013] In a possible implementation, the training process of the lower guard plate detection model is as follows:

[0014] Collect image data samples of the underbody guards of different vehicles under various lighting, weather conditions and driving speeds, including annotated images of normal and damaged states;

[0015] Preprocessing each lower guard plate image data sample, wherein the preprocessing includes: grayscale, denoising and contrast enhancement;

[0016] The preprocessed lower guard plate image data samples are input into the pre-trained target detection model for training, and the optimization algorithm is used to tune the hyperparameters. After reaching the preset training cutoff condition, the trained lower guard plate detection model is obtained.

[0017] In a possible implementation, the multiple lighting and weather conditions include: sunny days, cloudy days, nighttime and rainy days.

[0018] In a possible implementation, the target detection model adopts the YOLOv11 target detection model.

[0019] In a possible implementation, the optimization algorithm adopts a stochastic gradient descent SGD algorithm or an adaptive moment estimation Adam algorithm.

[0020] In a possible implementation, the camera module includes: a camera and a compensation unit;

[0021] The compensation unit is used to perform exposure and lighting compensation on the camera under various lighting and weather conditions.

[0022] In a possible implementation, the camera module further includes: a photoelectric sensor;

[0023] The photoelectric sensor is used to trigger the camera and the compensation unit to operate when a vehicle passes by the photoelectric sensor.

[0024] In a second aspect, an embodiment of the present application provides a method for detecting damage to a lower guard plate of an electric vehicle, comprising:

[0025] Acquire vehicle information and lower guard plate images of electric vehicles;

[0026] Perform damage detection on the lower guard plate image based on the trained lower guard plate detection model to obtain a detection result;

[0027] If the detection result shows that the lower guard plate of the electric vehicle is damaged, a warning message is sent to the corresponding vehicle or vehicle management system, and the warning message includes vehicle information, damage location and damage degree.

[0028] In a possible implementation, before performing damage detection on the lower guard plate image based on the trained lower guard plate detection model, the method further includes:

[0029] The lower guard plate image is preprocessed, and the preprocessing includes: grayscale, denoising and contrast enhancement.

[0030] Compared with the prior art, the electric vehicle underguard plate damage detection system provided in the present application includes at least one camera module, an image processing module and an early warning module; the at least one camera module is respectively connected to the image processing module, and the image processing module is connected to the early warning module; the camera module is installed at a fixed position under the road, and when the electric vehicle passes over the camera module, the image of the underguard plate of the electric vehicle is captured, and the image of the underguard plate is sent to the image processing module; the image processing module is used to perform damage detection on the underguard plate image based on a trained underguard plate detection model, and when the detection result is that the underguard plate of the electric vehicle is damaged, the damaged position and the degree of damage are sent to the early warning module; the early warning module is used to obtain vehicle information of the electric vehicle, and when receiving the detection result that the underguard plate of the electric vehicle is damaged, a warning message is sent to the corresponding vehicle or vehicle management system, and the warning message includes vehicle information, damaged position and degree of damage. Compared with the existing technology, the present application can perform real-time damage detection on the lower guard plate of electric vehicles based on a trained lower guard plate detection model, and promptly issue an alarm when the lower guard plate is found to be damaged, thereby preventing overheating or spontaneous combustion of electric vehicle batteries caused by direct or indirect damage to the liquid cooling plate due to damage to the lower guard plate. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present application. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0032] Figure 1 A structural schematic diagram of a system for detecting damage to an underbody guard plate of an electric vehicle provided in an embodiment of the present application is shown;

[0033] Figure 2 A schematic diagram of installing a camera module provided in an embodiment of the present application is shown;

[0034] Figure 3 A flow chart of detecting damage to the lower guard plate of an electric vehicle based on the lower guard plate damage detection system of the present application is shown;

[0035] Figure 4 A flow chart of a method for detecting damage to a lower guard plate of an electric vehicle provided in an embodiment of the present application is shown;

[0036] Figure 5 A schematic diagram of an electronic device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0037] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0038] It should be noted that, unless otherwise specified, the technical terms or scientific terms used in this application should have the common meanings understood by technicians in the field to which this application belongs.

[0039] In addition, the terms "first" and "second" etc. are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or devices.

[0040] At present, there is no good automated detection solution for outdoor electric vehicle underbody guard plate inspection. The manual visual method has a long detection cycle, is time-consuming and labor-intensive, and traditional detection methods such as pressure monitoring and thermal imaging are costly, inaccurate, and have limited application scenarios. Therefore, this application proposes to install a camera under the road surface to collect images of the electric vehicle chassis in real time, and use a deep learning target detection algorithm to automatically identify damage to the underbody guard plate. This application aims to replace traditional manual inspection methods, reduce detection time and cost, improve detection efficiency, and be able to complete real-time safety monitoring of the underbody guard plate during vehicle driving.

[0041] Please refer to Figure 1 , Figure 1 1 is a schematic diagram of a structure of a system for detecting damage to the lower guard plate of an electric vehicle provided in an embodiment of the present application, the detection system comprising: at least one camera module 10, an image processing module 20 and an early warning module 30; the at least one camera module 10 is respectively connected to the image processing module 20, and the image processing module 20 is connected to the early warning module 30. The at least one camera module 10 can be set on different lanes, or two or more camera modules 10 can be set at different positions in the same lane, which is not limited in the present application.

[0042] Each camera module 10 is installed at a fixed position under the road. When the electric vehicle passes over the camera module, the image of the lower guard plate of the electric vehicle is captured and sent to the image processing module 20.

[0043] like Figure 2The figure shows an installation diagram of a camera module 10. Specifically, the camera module 10 includes: a camera and a compensation unit; the compensation unit is used to perform exposure and illumination compensation on the camera under various lighting and weather conditions. The compensation unit can be a light source.

[0044] Specifically, the camera can be installed under the road or at the bottom of the detection site to ensure that a clear chassis image can be obtained, especially in the area of ​​the lower guard plate. The camera has a high resolution and high frame rate to deal with the problem of image blur when the vehicle is driving at high speed. In addition, in order to adapt to different lighting environments, the camera is equipped with automatic exposure and lighting compensation functions to ensure high-quality images in day and night and in all weather conditions.

[0045] like Figure 2 As shown, the camera module 10 further includes: a photoelectric sensor; the photoelectric sensor is used to trigger the camera and the compensation unit to work when a vehicle passes by the photoelectric sensor.

[0046] The image processing module 20 is used to perform damage detection on the lower guard plate image based on the trained lower guard plate detection model, and when the detection result shows that the lower guard plate of the electric vehicle is damaged, the damage location and damage degree are sent to the early warning module;

[0047] The lower guard plate detection model can be built based on the YOLO (You Only Look Once) target detection algorithm. The YOLO target detection algorithm is a real-time target detection algorithm with an end-to-end structure that can simultaneously complete target positioning and classification in the process of one-time image processing. The technical principle of YOLO is based on the convolutional neural network (CNN). Its feature extraction layer can identify the bounding box of the lower guard plate area from the image and classify the area as "normal" or "damaged". YOLO can ensure efficient recognition of different types of damage to the lower guard plate (such as cracks, dents, perforations, etc.) by performing detection on feature maps of different scales. YOLO uses convolutional layers and fully connected layers to extract image features, and uses regression methods to directly predict the category and position coordinates of the object. Its notable feature is that it can quickly process a large number of images, which is suitable for high-frequency detection scenarios in electric vehicle driving.

[0048] When damage is detected in the lower guard plate, the damage area ratio and damage depth estimation can be performed: the ratio of the detected damaged area to the entire image area is calculated. The Sobel operator is used to calculate the gradient amplitude of the damaged area to approximately measure the depth feature. The gradient represents the rate and direction of change of pixel brightness in the image. If the gradient value is high, it means that the surface of the area has changed dramatically, which may be deep damage. Set the threshold to classify the damage. The threshold can be adjusted according to the data distribution of different actual scenes.

[0049] In some embodiments, before performing damage detection on the lower guard plate image based on a trained lower guard plate detection model, the image processing module 20 is further used to preprocess the lower guard plate image, wherein the preprocessing includes graying, denoising and contrast enhancement.

[0050] Specifically, in order to improve the accuracy of the under guard plate detection model, the image captured by the camera is first preprocessed. The main preprocessing steps include grayscale (reducing color information interference), denoising (eliminating noise interference in the image), contrast enhancement (highlighting the edge and details of the under guard plate), etc. These technologies ensure that the image still has high quality and clarity under complex background or environmental conditions.

[0051] This application adopts the YOLO target detection algorithm, combined with real-time image processing technology, to ensure that the system can quickly identify the deformation or damage of the lower guard plate during the driving of the electric vehicle. The core of the detection system is to install the camera under the road surface, continuously capture images of the chassis of the moving vehicle, transmit them to the image processing module, and analyze them by the YOLO algorithm. The system accurately locates the lower guard plate area through the bounding box and determines whether there is damage in the area. In addition, the system provides data feedback and evaluation of the detection results, and through continuous model training and optimization, improves the accuracy and real-time performance of the detection, ensuring its stability and robustness in practical applications.

[0052] The warning module 30 is used to obtain vehicle information of the electric vehicle. When receiving the detection result that the lower guard plate of the electric vehicle is damaged, it sends a warning message to the corresponding vehicle or vehicle management system, and the warning message includes vehicle information, damage location and damage degree. The vehicle information is used to identify the vehicle, and may include information such as license plate number, vehicle model, and vehicle body color.

[0053] In this application, once the lower guard plate is detected to be damaged or abnormal, the system will issue an alarm through the early warning module. The early warning module is connected to the background management system and can send alarm information to the vehicle driver, vehicle maintenance team or traffic monitoring center. The early warning information includes the license plate number, damage location and degree of damage, etc., which facilitates the rapid response of subsequent repairs or maintenance work. The early warning module also has data storage and log recording functions for later statistical analysis and fault tracing.

[0054] For ease of understanding, the present application also provides Figure 3 The flowchart shown is for detecting damage to the lower guard plate of an electric vehicle based on the system for detecting damage to the lower guard plate of the present application.

[0055] This application installs a camera under the road surface, uses artificial intelligence algorithms to monitor the vehicle chassis in real time and promptly identify potential underbody damage. Once underbody damage is detected, the system will promptly issue an early warning to indicate that the vehicle is at potential safety risk. The system can not only efficiently and in real time capture chassis images, but also automatically analyze the damage status to avoid human errors in manual inspections.

[0056] The training process of the lower guard plate detection model is described in detail below. The training process of the lower guard plate detection model is as follows:

[0057] Collect lower guard plate image data samples of different vehicles under various lighting, weather conditions and driving speeds, including annotated images of normal and damaged states; specifically, the various lighting and weather conditions include: sunny days, cloudy days, nighttime and rainy days.

[0058] During the imaging and data collection process of the lower guard plate, image data under various conditions are collected for the lower guard plate of the electric vehicle. Experiments are conducted under different lighting conditions, different imaging devices, and shooting angles to obtain clear imaging data of the lower guard plate. To ensure the comprehensiveness of the data, the image set covers the normal state of the lower guard plate and various damage conditions (such as cracks, dents, breakage, etc.). By analyzing the imaging results, the imaging characteristics of the lower guard plate such as reflection, shadow, and noise under different conditions are studied. These studies provide basic theoretical support for the subsequent development of detection algorithms and ensure that the system can work normally under complex road conditions and lighting environments.

[0059] Each lower guard plate image data sample is preprocessed, and the preprocessing includes: grayscale, denoising, contrast enhancement, etc. These technologies ensure that the image still has high quality and clarity under complex background or environmental conditions.

[0060] The preprocessed lower guard plate image data samples are input into the pre-trained target detection model for training, and the optimization algorithm is used to tune the hyperparameters. After reaching the preset training cutoff condition, the trained lower guard plate detection model is obtained.

[0061] The damage of the lower guard plate usually manifests itself in different forms of features such as deformation and cracks. Using image processing technology, the image is first preprocessed (such as grayscale, denoising, contrast enhancement, etc.), and then the feature extraction of the lower guard plate area is performed through the target detection algorithm to separate the image features of normal and damaged lower guard plates. In order to improve the recognition accuracy, a deep learning network is used to train the specific features of the damage of the lower guard plate (such as the edge of the crack, the texture of the damaged area, etc.) to ensure that the algorithm can accurately extract the damage features of the lower guard plate under complex backgrounds.

[0062] Specifically, the target detection model adopts the YOLOv11 target detection model. The beneficial effects of the YOLO target detection algorithm adopted in the embodiment of the present application are as follows:

[0063] 1. Real-time and automation: Using the fast processing capability of the artificial intelligence target detection algorithm YOLO, the system can automatically and in real time identify whether the lower guard plate is damaged while the vehicle is driving. This avoids the drawbacks of traditional methods that require the vehicle to stop for inspection or regular maintenance, greatly improving the efficiency and practicality of detection.

[0064] 2. Image-based non-contact detection: Different from the traditional sensor detection method, the system realizes non-contact detection through image acquisition and analysis, which not only reduces the detection cost, but also avoids the limitation of sensor installation position.

[0065] 3. Application of deep learning algorithms: The YOLO algorithm has strong robustness and can adapt to different lighting conditions and the complex structure of the vehicle chassis, enabling the system to work stably under a variety of environmental conditions. At the same time, YOLO's end-to-end design improves detection accuracy and speed, ensuring that damage to the lower guard plate can be identified in a timely manner.

[0066] The pre-trained YOLO model can be further fine-tuned through transfer learning to enable it to accurately detect the damage characteristics of the lower guard plate. In order to improve the detection efficiency of the model, this application also uses an optimization algorithm to tune hyperparameters to ensure that the model achieves the best balance between detection accuracy and real-time performance. The optimization algorithm uses the stochastic gradient descent SGD algorithm or the adaptive moment estimation Adam algorithm.

[0067] In order to solve the complexity of lower guard plate detection, this application introduces the YOLO target detection algorithm. As one of the representative algorithms in the field of deep learning, YOLO can complete multi-target detection of images in a relatively short time, and is suitable for real-time analysis of chassis images during driving of electric vehicles. By utilizing the powerful feature extraction capability of the YOLO algorithm, the system can quickly and accurately identify the deformation or damage of the lower guard plate, and issue an early warning when necessary to prevent the potential risk of spontaneous combustion caused by overcharging and high temperature. This method not only improves the detection efficiency, but also reduces the detection cost. It provides an innovative solution for the safety of electric vehicles. It has important practical value, can effectively improve the safety of electric vehicles, and prevent high-temperature spontaneous combustion accidents caused by damage to the lower guard plate.

[0068] The following are the experimental process and results of the electric vehicle underbody guard damage detection system provided in this application.

[0069] (I) Experimental process

[0070] 1. Experimental environment and equipment settings:

[0071] The experiment was conducted in a simulated road scene. Four high-resolution cameras were installed under the road surface, distributed in different lanes. The shutter speed of the camera was set to 1 / 1000 second to meet the imaging needs of vehicles traveling at high speeds. Each camera was connected to an image processing server, which used the YOLOv11 target detection model for real-time analysis. In order to test the performance of the system under various conditions, the experiment selected a variety of lighting and weather conditions, including sunny, cloudy, night and rainy days.

[0072] 2. Data collection:

[0073] Five electric vehicles were selected during the data collection phase. The lower guard plate conditions of the experimental vehicles covered three conditions: normal, slightly damaged, and severely damaged. The types of damage to the lower guard plate include cracks, dents, and perforations. Under each experimental condition (such as sunny days and nighttime), the vehicle drove through the detection area at different speeds (50km / h to 100km / h), and the system automatically captured the vehicle chassis image. A total of 1,500 image data were collected, including 500 normal lower guard plate images, 500 slightly damaged images, and 500 severely damaged images.

[0074] 3. Image processing and detection:

[0075] After data collection is completed, the system automatically preprocesses each image, including grayscale, denoising, and contrast enhancement. The preprocessed image is input into the YOLOv11 target detection model, which automatically identifies the damaged area of ​​the lower guard plate and generates detection results, including the damage type and damaged area of ​​the lower guard plate.

[0076] 4. Experimental indicators:

[0077] In order to evaluate the performance of the system, this experiment used the following evaluation indicators:

[0078] Accuracy: The ratio of the detection system correctly identifying the status of the lower guard plate.

[0079] Recall rate: The proportion of all damaged lower guard plates that the system successfully detects.

[0080] Precision: The proportion of damaged lower guard plates detected by the system that are actually damaged.

[0081] F1-Score: The weighted average of precision and recall, used to measure the overall performance of the system.

[0082] Detection speed: the average time from image acquisition to generating detection results.

[0083] (II) Experimental results

[0084] After multiple rounds of experiments, the performance of the system in different environments and driving speeds is summarized as follows:

[0085] 1. Detection accuracy:

[0086] The system achieved the highest accuracy of 98.5% on sunny days, with stable results for normal lower guards and mild damage. At night and in rainy weather, the system maintained an accuracy of 94.7% despite insufficient light and rain. The system was particularly effective for severely damaged lower guards, with an accuracy of 99.2%.

[0087] 2. Recall and precision:

[0088] The experimental results show that the system's recall rate for damaged lower guard plates is 96.8%, indicating that most damaged lower guard plates are successfully detected. At the same time, the precision rate reaches 97.3%, indicating that the system's recognition of damage is relatively accurate and the false alarm rate is low.

[0089] 3.F1 score:

[0090] Under different experimental conditions, the system's F1 score remained above 97.0, especially in daytime and well-lit environments, where the F1 score reached 98.0, indicating that the system achieved a good balance between precision and recall.

[0091] 4. Detection speed:

[0092] When the vehicle passes through the detection area at different speeds, the system's average detection time is 0.25 seconds. Even if the vehicle speed reaches 100km / h, the system can still complete the detection and generate results in no more than 0.3 seconds, which fully demonstrates the system's real-time processing capabilities.

[0093] 5. Influence of environmental factors:

[0094] The experiment also analyzed the impact of different lighting and weather conditions on system performance. Under sunny conditions, the image clarity is high, and the system's detection accuracy and speed are the best. At night, although the system's accuracy rate drops slightly, with the help of automatic exposure adjustment and infrared fill light function, the system still maintains a high detection effect. Under rainy conditions, raindrops and road water have a certain impact on image quality, resulting in a slight increase in the false alarm rate, but the overall impact is not significant, and the detection accuracy rate is still over 94%.

[0095] Overall, the experimental results prove that the electric vehicle underbody panel damage detection system proposed in this application has high accuracy and real-time performance, especially when driving at high speeds, it can still effectively detect underbody panel damage. Whether in normal environments or complex lighting conditions, the system exhibits strong robustness. The experimental results show that this system can be used as an effective tool for electric vehicle underbody panel damage detection and is suitable for practical application scenarios.

[0096] The excellent performance of this system is due to the following aspects:

[0097] 1. The efficiency of the artificial intelligence target detection algorithm YOLO

[0098] The key to the good results of the system lies in the use of the YOLO target detection algorithm. As an end-to-end target detection algorithm, YOLO can simultaneously complete object positioning and classification in a single forward pass, greatly improving the detection speed. In the experiment, the YOLO algorithm performed well, especially when the vehicle was traveling at a high speed, it could still maintain a detection time of 0.25 seconds, ensuring the real-time performance of the system. This provides strong technical support for the detection of underbody damage during vehicle driving.

[0099] 2. Application of image preprocessing and data enhancement technology

[0100] The success of the experiment also benefited from effective image preprocessing and data enhancement technology. Under different environmental conditions, the system improved the image quality through grayscale, denoising, contrast enhancement and other processing methods, especially improving the imaging effect of the lower guard plate under complex lighting. In addition, data enhancement technology makes the YOLO model more robust by simulating scenes under a variety of different angles and lighting conditions, thereby maintaining high detection accuracy and recall rate under various experimental conditions.

[0101] 3. Test coverage in multiple environments

[0102] The experimental design takes into account a variety of lighting and weather conditions (such as sunny days, nighttime, rainy days, etc.), and collects a large number of images under different conditions to ensure that the model can adapt to complex practical application environments. This diverse data collection and experimental design helps the system maintain stable detection performance in a variety of scenarios, especially in daytime and sunny environments, with a system accuracy rate of up to 98.5%, indicating that the model has good generalization ability.

[0103] 4. Optimization of Image Acquisition

[0104] The camera module plays a key role in the system's performance. Through automatic exposure adjustment and fill light technology, the system is able to maintain high image quality at night and in low light conditions. This technology effectively compensates for the image quality degradation caused by changes in ambient light, allowing the system's detection accuracy to remain above 94% at night and in rainy days.

[0105] In the future, with further optimization of models and systems, detection accuracy and applicable scenarios are expected to be further expanded, providing efficient safety protection for more electric vehicles.

[0106] The present application also provides a method for detecting damage to the lower guard plate of an electric vehicle corresponding to the detection system provided in the above-mentioned embodiment. Figure 4 As shown, the method includes:

[0107] S101, obtaining vehicle information and a lower guard plate image of the electric vehicle;

[0108] S102, performing damage detection on the lower guard plate image based on the trained lower guard plate detection model to obtain a detection result;

[0109] S103. If the detection result shows that the lower guard plate of the electric vehicle is damaged, a warning message is sent to the corresponding vehicle or vehicle management system, and the warning message includes vehicle information, damage location and damage degree.

[0110] The training process of the lower guard plate detection model refers to the above embodiment and will not be repeated here.

[0111] In a possible implementation, before performing damage detection on the lower guard plate image based on the trained lower guard plate detection model in step S102, the method further includes:

[0112] The lower guard plate image is preprocessed, and the preprocessing includes: grayscale, denoising and contrast enhancement.

[0113] An embodiment of the present application also provides an electronic device corresponding to the method provided in the aforementioned embodiment, and the electronic device may be a laptop computer, a desktop computer, a server, etc., to execute the above-mentioned electric vehicle underbody panel damage detection method.

[0114] Please refer to Figure 5 , which shows a schematic diagram of an electronic device provided by some embodiments of the present application. Figure 5As shown, the electronic device 40 includes: a processor 200, a memory 201, a bus 202 and a communication interface 203, and the processor 200, the communication interface 203 and the memory 201 are connected via the bus 202; the memory 201 stores a computer program that can be run on the processor 200, and when the processor 200 runs the computer program, it executes the electric vehicle underbody panel damage detection method provided in any of the aforementioned embodiments of the present application.

[0115] The memory 201 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk memory. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 203 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used.

[0116] The bus 202 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The memory 201 is used to store programs, and the processor 200 executes the programs after receiving the execution instructions. The electric vehicle lower guard plate damage detection method disclosed in any implementation of the above-mentioned embodiment of the present application can be applied to the processor 200, or implemented by the processor 200.

[0117] The processor 200 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 200. The above processor 200 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a readily available programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor to be executed, or the hardware and software modules in the decoding processor can be executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 201, and the processor 200 reads the information in the memory 201 and completes the steps of the above method in combination with its hardware.

[0118] The electronic device provided in the embodiment of the present application and the method for detecting damage to the lower guard plate of an electric vehicle provided in the embodiment of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, operated or implemented therein.

[0119] An embodiment of the present application also provides a computer-readable storage medium corresponding to the electric vehicle lower guard plate damage detection method provided in the aforementioned embodiment, on which a computer program (i.e., a program product) is stored. When the computer program is run by a processor, it will execute the electric vehicle lower guard plate damage detection method provided in any of the aforementioned embodiments.

[0120] It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other optical or magnetic storage media, which are not listed here one by one.

[0121] The computer-readable storage medium provided in the above-mentioned embodiments of the present application and the electric vehicle underbody guard plate damage detection method provided in the embodiments of the present application are based on the same inventive concept and have the same beneficial effects as the methods adopted, run or implemented by the application programs stored therein.

[0122] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and they should all be included in the scope of the claims and specification of the present application.

Claims

1. A system for detecting damage to the lower guard plate of an electric vehicle, characterized in that: include: At least one camera module, an image processing module and an early warning module; the at least one camera module is connected to the image processing module respectively, and the image processing module is connected to the early warning module; The camera module is installed at a fixed position below the road. When the electric vehicle passes over the camera module, the image of the lower guard plate of the electric vehicle is captured, and the image of the lower guard plate is sent to the image processing module. The image processing module is used to perform damage detection on the lower guard plate image based on the trained lower guard plate detection model, and when the detection result shows that the lower guard plate of the electric vehicle is damaged, the damage location and damage degree are sent to the early warning module; The early warning module is used to obtain vehicle information of the electric vehicle. When receiving the detection result that the lower guard plate of the electric vehicle is damaged, it sends early warning information to the corresponding vehicle or vehicle management system. The early warning information includes vehicle information, damage location and damage degree.

2. The electric vehicle underbody damage detection system according to claim 1, characterized in that: Before performing damage detection on the lower guard plate image based on the trained lower guard plate detection model, the image processing module is further used to: The lower guard plate image is preprocessed, and the preprocessing includes: grayscale, denoising and contrast enhancement.

3. The electric vehicle underbody damage detection system according to claim 1, characterized in that: The training process of the lower guard plate detection model is as follows: Collect image data samples of the underbody guards of different vehicles under various lighting, weather conditions and driving speeds, including annotated images of normal and damaged states; Preprocessing each lower guard plate image data sample, wherein the preprocessing includes: grayscale, denoising and contrast enhancement; The preprocessed lower guard plate image data samples are input into the pre-trained target detection model for training, and the optimization algorithm is used to tune the hyperparameters. After reaching the preset training cutoff condition, the trained lower guard plate detection model is obtained.

4. The electric vehicle underbody damage detection system according to claim 3, characterized in that: The various lighting and weather conditions include: sunny days, cloudy days, nighttime and rainy days.

5. The electric vehicle underbody damage detection system according to claim 3, characterized in that: The target detection model adopts the YOLOv11 target detection model.

6. The electric vehicle underbody panel damage detection system according to claim 3, characterized in that: The optimization algorithm adopts a stochastic gradient descent SGD algorithm or an adaptive moment estimation Adam algorithm.

7. The electric vehicle underbody damage detection system according to claim 1, characterized in that: The camera module includes: a camera and a compensation unit; The compensation unit is used to perform exposure and lighting compensation on the camera under various lighting and weather conditions.

8. The electric vehicle underbody damage detection system according to claim 7, characterized in that: The camera module further includes: a photoelectric sensor; The photoelectric sensor is used to trigger the camera and the compensation unit to operate when a vehicle passes by the photoelectric sensor.

9. A method for detecting damage to the lower guard plate of an electric vehicle, characterized in that: include: Acquire vehicle information and lower guard plate images of electric vehicles; Perform damage detection on the lower guard plate image based on the trained lower guard plate detection model to obtain a detection result; If the detection result shows that the lower guard plate of the electric vehicle is damaged, a warning message is sent to the corresponding vehicle or vehicle management system, and the warning message includes vehicle information, damage location and damage degree.

10. The electric vehicle lower guard plate damage detection method according to claim 9, characterized in that: Before performing damage detection on the lower guard plate image based on the trained lower guard plate detection model, the method further includes: The lower guard plate image is preprocessed, and the preprocessing includes: grayscale, denoising and contrast enhancement.

Images