Vehicle body scratch recognition and acquisition equipment

Abstract

The invention discloses vehicle body scratch recognition and acquisition equipment, relates to the field of machine vision and automation, and aims to solve the problems of low efficiency, insufficient accuracy, poor environmental adaptability and limited data processing capability of traditional vehicle body scratch detection. The equipment comprises an image acquisition unit, an image processing unit, a data storage and analysis unit, a user interface, an automatic control unit, an integrated network interface and an aluminum alloy waterproof and dustproof main body. The image acquisition unit adopts a high-resolution camera and an LED cross illumination system to adapt to different indoor and outdoor light conditions; the automatic control unit triggers full-process automatic detection through a ground sensor, and a multi-view scene supports all-directional synchronous detection of a vehicle body. The method is suitable for multi-scene deployment of various vehicle types, indoor workshops, outdoor parking lots and the like, provides accurate data support for vehicle maintenance, insurance claim settlement, second-hand vehicle valuation and the like, and remarkably improves the detection efficiency and practicability.

Description

Technical Field

[0001] This invention relates to the field of machine vision and automation, and in particular to a vehicle body scratch recognition and acquisition device. Background Technology

[0002] Vehicle scratches are a common form of damage during vehicle use, directly impacting the vehicle's appearance and value. Traditional scratch detection relies primarily on manual visual inspection, a method that is not only inefficient but also susceptible to human error, leading to missed or incorrect detections. With the rapid development of the automotive industry and increasing consumer demands for vehicle aesthetics, there is an urgent need for efficient and accurate scratch recognition and acquisition equipment.

[0003] Currently, there are some vehicle scratch recognition systems on the market based on image processing and machine vision technologies, but these systems still have some problems. For example, outdoor dust and changes in lighting can cause image quality to degrade, thus affecting the accuracy of scratch recognition. In addition, traditional scratch recognition systems often have insufficient recognition capabilities when processing complex scratch features. At the same time, existing systems also have certain limitations in the acquisition, storage, and analysis of scratch data, failing to meet the growing market demand.

[0004] To address the aforementioned problems, this invention patent proposes a novel vehicle body scratch recognition and acquisition device, aiming to achieve efficient and accurate identification and acquisition of vehicle body scratches through advanced technology. This device employs a high-resolution camera and intelligent algorithms to quickly capture images of the vehicle body surface and achieves accurate scratch identification through image preprocessing, feature extraction, and scratch detection. Simultaneously, the device also features data recording and analysis capabilities, providing strong technical support for vehicle repair, insurance claims, and other fields. Summary of the Invention

[0005] To address the technical problems of low efficiency, insufficient accuracy, limited data processing capabilities, and poor environmental adaptability in existing vehicle body scratch detection systems, this invention provides a vehicle body scratch recognition and acquisition device. Through modular design and integration with advanced technologies, it achieves efficient scratch recognition, accurate acquisition, and intelligent analysis.

[0006] This invention is achieved using the following technical solution: a vehicle body scratch recognition and acquisition device, comprising an image acquisition unit, an image processing unit, a data storage and analysis unit, a user interface, and an automation control unit; the image acquisition unit is used to capture images of the vehicle body surface, the image processing unit is used to preprocess, extract features, and identify scratches from the acquired images, the data storage and analysis unit is used to store scratch data and perform statistical analysis, the user interface is used for device operation and data display, and the automation control unit is used to control the device operation and image acquisition process; the device also integrates a network interface module with the main body, and each unit is electrically connected through internal wiring to collaboratively complete the entire process of scratch recognition and data acquisition.

[0007] As a further optimization of the present invention, the image acquisition unit includes a high-resolution industrial camera using a CMOS sensor and an LED lighting system. The high-resolution camera supports focus adjustment and exposure adjustment functions, and the image output format is JPEG. For outdoor scenarios, an IP67 waterproof and dustproof camera is used, with a lens equipped with a hydrophobic coating and a sunshade. The LED lighting system is symmetrically distributed on both sides of the camera, each group containing multiple high-brightness surface-mount LEDs, using a cross-illumination angle. It has a built-in light sensor module, and the brightness can be steplessly adjusted via an automated control unit. In outdoor scenarios, it can adaptively adjust in real time according to the ambient light intensity. The effect is that, through the high-resolution camera and optimized LED lighting design, it ensures the capture of high-definition, non-reflective, and shadowless images of the vehicle body surface under different indoor and outdoor lighting conditions. The design of the hydrophobic coating, sunshade, and light sensor module further improves the imaging stability of the equipment in complex environments such as outdoor rain and strong light, providing a high-quality image foundation for subsequent scratch recognition.

[0008] As a further optimization of the present invention, the image processing unit adopts either a high-performance ARM or a deep learning chip as the main embedded processor chip, and is equipped with a DSP as an auxiliary chip to carry out the computational tasks of image preprocessing, feature extraction, and scratch recognition. The deep learning chip can be an NPU chip; for outdoor scenarios, a more reliable DSP chip is preferred. This unit is divided into an image preprocessing module, a feature extraction module, and a scratch recognition module. The image preprocessing module uses a mean filtering algorithm to denoise the image, enhances image contrast through gamma correction, and converts the color image into an 8-bit grayscale image. The feature extraction module extracts image edge features based on the Canny edge detection algorithm, analyzes scratch texture features using the gray-level co-occurrence matrix, and simultaneously collects the difference between the scratch area and the vehicle's base color, as well as contour shape parameters (length, width, area, curvature), forming a multi-dimensional feature vector. A branch recognition algorithm is added for intersecting scratches, and multiple intersecting scratch parameters are distinguished through contour segmentation technology. The scratch recognition module incorporates a lightweight deep learning model optimized based on the MobileNet architecture, trained with over 100,000 vehicle scratch samples, and adds a scratch depth grading function for outdoor scenarios. By combining dedicated chips with optimized algorithms, image processing efficiency is significantly improved. The extraction of multi-dimensional feature vectors ensures the comprehensiveness of scratch features. The lightweight deep learning model reduces hardware resource consumption while ensuring an accuracy rate of ≥98%. The cross-scratch recognition algorithm and deep grading function further expand the device's ability to handle complex scratches, providing more accurate technical basis for maintenance assessment.

[0009] As a further optimization of this invention, the data storage and analysis unit includes a 512GB SATA interface solid-state drive and a data analysis module. An AES-256 encryption module is added for outdoor scenarios. The solid-state drive employs a partitioned storage strategy, with image files named according to "acquisition time - vehicle identification code." Recognition results are stored in JSON format, containing information such as scratch ID, location coordinates, size parameters, and recognition confidence level, supporting local data retention for more than one year. The data analysis module incorporates statistical algorithms to statistically analyze scratch detection quantity, scratch type distribution, and high-frequency scratch locations daily / weekly / monthly, generating statistical reports such as bar charts and pie charts. A new full-vehicle scratch statistics function is added for multi-view scenarios, generating reports on the number of scratches on the entire vehicle, scratch distribution in each direction, and severity percentage. The large-capacity solid-state drive meets long-term data storage needs, standardized naming and data formats facilitate rapid retrieval, and the encryption module ensures the storage security of image data and recognition results. Diverse statistical analysis functions not only provide data support for repair plan formulation and insurance risk assessment but also adapt to full-vehicle inspection scenarios, providing comprehensive data references for used car valuation and comprehensive repair planning.

[0010] As a further optimization of the present invention, the user interface includes a capacitive touchscreen and physical buttons. The touchscreen can display the device's operating status (standby / acquisition / processing / fault), real-time acquired images, scratch recognition and marking results, statistical reports, etc., and supports users to manually zoom in to view local images and export data files. In multi-view scenarios, it supports split-screen viewing of scratch images from various directions or synthesizing a full-vehicle scratch distribution diagram. The physical buttons include a power button (controlling device start and stop), an emergency stop button (cutting off device operation in abnormal situations), and an acquisition confirmation button (supporting manual triggering of image acquisition, adapting to reshoot scenarios). The high-definition display and interactive functions of the capacitive touchscreen make data viewing more intuitive, the physical button settings meet the needs of rapid operation and emergency handling, and the split-screen viewing and full-vehicle diagram functions improve the ease of operation in multi-view detection scenarios. The overall design takes into account both automated detection and manual intervention needs, improving the user experience in different usage scenarios.

[0011] As a further optimization of this invention, the automated control unit includes two ground sensor units and one microcontroller. In multi-view scenarios, four additional ground sensor units are added, each corresponding to a detection area around the vehicle body. The ground sensor units are embedded at the entrance, exit, and corresponding positions in each direction of the detection area, generating electrical signals by sensing the vehicle's metal chassis. The microcontroller receives the sensor signals. When the vehicle enters the entrance ground sensor area, it triggers a high-resolution camera to continuously capture 8-11 images (covering the vehicle body detection surface). In multi-view scenarios, it coordinates and controls the four image acquisition units to work synchronously. When the vehicle exits the exit ground sensor area, it controls the camera to stop acquiring images. The ground sensor triggering mechanism achieves fully automated image acquisition without manual intervention. The continuous shooting design ensures comprehensive coverage of the vehicle body detection surface. The sensor expansion and synchronous control functions in multi-view scenarios prevent missed detections of vehicle body orientations, significantly improving detection efficiency and completeness, and adapting to different needs from single-position detection to full-range vehicle detection.

[0012] As a further optimization of this invention, the device integrates a network interface module, including one gigabit Ethernet wired network interface and one Wi-Fi module. Network transmission uses the HTTPS protocol, supporting wired or wireless communication connections between the device and remote servers, maintenance management systems, or insurance claims platforms. The advantages are: dual network connection methods adapt to the network deployment needs of different installation environments; gigabit Ethernet ensures high-speed transmission of large amounts of data (images + recognition results); the Wi-Fi module improves the flexibility of device deployment; and the HTTPS protocol ensures the security and integrity of remote data transmission, enabling real-time sharing and remote viewing of recognition results, thus meeting the collaborative work needs of maintenance, insurance, and other related fields.

[0013] As a further optimization of the present invention, the outer shell of the main body of the equipment is made of aluminum alloy through die casting, and the surface is anodized and then sprayed with a waterproof and dustproof coating, achieving a protection level of IP65. For outdoor applications, the shell is upgraded to a thicker aluminum alloy material with an added heat insulation layer, adapting to an outdoor temperature range of -40℃ to 85℃. The main body of the equipment is equipped with an adjustable-height mounting bracket, which is upgraded to a movable bracket with wheels (the wheels are equipped with a braking device) for outdoor applications. The adjustable height range of the bracket can accommodate different vehicle types such as sedans, SUVs, trucks, and off-road vehicles. The waterproof, dustproof, and heat-insulating design of the aluminum alloy shell significantly improves the environmental adaptability of the equipment, resisting harsh conditions such as workshop dust, outdoor rain, and extreme temperatures, extending the service life of the equipment. The adjustable and movable bracket design allows the equipment to flexibly adapt to the detection height requirements of different vehicle types and the installation and deployment requirements of different scenarios, reducing the limitations of equipment use and expanding its application scenarios.

[0014] As a further optimization of the present invention, the embedded processor of the image processing unit can be flexibly selected according to the application scenario. Among them, the deep learning chip has stronger parallel computing capabilities, adapting to the parallel processing requirements of complex scratches (cross scratches, mixed deep and shallow scratches). The effect is that the flexible selection of the processor allows the device to be specifically adapted to different detection scenarios, and the parallel computing advantage of the deep learning chip effectively shortens the processing time of complex scratches, ensuring that even in high-efficiency outdoor detection scenarios, recognition results can be quickly output even when faced with complex scratch features, balancing detection accuracy and processing efficiency.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention significantly improves detection efficiency and achieves full-process automation. It triggers image acquisition via a ground sensor, eliminating the need for manual intervention. The entire process of shooting, processing, and identifying a vehicle is automatically completed as it enters the field. 8-11 images are captured continuously to ensure complete coverage. Four acquisition units work simultaneously in multi-view scenarios to avoid missed detections due to varying azimuth. Compared to traditional manual visual inspection, this invention not only completely eliminates reliance on human operation and significantly reduces the inspection time per vehicle, but also adapts to batch inspection needs. It effectively solves the pain points of low efficiency and high miss rate associated with traditional methods, making it suitable for high-frequency inspection scenarios such as repair shops, parking lots, and vehicle inspections.

[0016] 2. This invention boasts high recognition accuracy and strong ability to handle complex scratches. Based on a lightweight deep learning model trained with over 100,000 samples, the scratch recognition accuracy is ≥98%, accurately distinguishing scratches from interfering factors such as paint texture and stains. Through multi-dimensional feature extraction and cross-scratch branch recognition algorithms, it can accurately capture parameters of linear, dotted, irregular, and cross-shaped scratches. The newly added depth grading function for outdoor scenarios provides a quantitative basis for repair cost assessment, completely solving the problems of insufficient ability and poor accuracy in complex scratch recognition of traditional systems, and providing reliable data support for subsequent business operations.

[0017] 3. This invention boasts wide environmental adaptability and flexible deployment. The main body of the equipment utilizes an aluminum alloy shell with a waterproof and dustproof coating, allowing it to withstand extreme temperatures ranging from -40℃ to 85℃ and the corrosive effects of rain and dust after upgrades for outdoor use. The LED lighting system includes a light-sensing module, and the camera is equipped with a hydrophobic coating and a sunshade, ensuring clear imaging under various indoor and outdoor lighting conditions. Adjustable or movable brackets are compatible with various vehicle types, including sedans, SUVs, and trucks, enabling flexible deployment in multiple scenarios such as indoor workshops, outdoor parking lots, and roadside assistance, overcoming the limitations of traditional equipment in terms of poor environmental adaptability and limited vehicle compatibility. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the working process of the device of the present invention; Figure 2 This is a schematic diagram of the ground sensor triggering structure of the present invention; Figure 3 This is a schematic diagram of the device of the present invention. Detailed Implementation

[0019] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments. It should be noted that, without conflict, the various embodiments or technical features described below can be arbitrarily combined to form new embodiments.

[0020] Example 1: like Figure 1 (Equipment workflow diagram) and Figure 2 (Diagram of ground sensor triggering structure) As shown, the vehicle body scratch recognition and acquisition device in this embodiment includes a main body, an image acquisition unit, an image processing unit, a data storage and analysis unit, a user interface, an automation control unit, and an integrated network interface module. Each unit is electrically connected through internal circuitry to collaboratively complete the scratch recognition and data acquisition functions.

[0021] 1.1.1 Main body of the equipment The main body of the equipment is made of aluminum alloy through die casting. After anodizing, the surface is coated with a waterproof and dustproof coating, which can effectively resist the corrosion of internal components by workshop dust and minor water stains. The bottom of the equipment is equipped with an adjustable mounting bracket, which can be fixed to the ground or workbench with bolts to meet the testing height requirements of different vehicle models (cars, SUVs, etc.).

[0022] 1.1.2 Image Acquisition Unit The image acquisition unit, which includes a high-resolution industrial camera and an LED lighting system, is fixed to a modular mounting base at the front of the main body of the device. The high-resolution camera uses a CMOS sensor and supports focus adjustment and exposure adjustment functions. The image output format is JPEG, ensuring fast capture of high-definition images of the vehicle surface. The LED lighting system is symmetrically distributed on both sides of the camera. Each group contains multiple high-brightness surface-mount LEDs and adopts a cross-illumination angle. The brightness can be steplessly adjusted by the automatic control unit to ensure uniform, shadow-free illumination of the vehicle surface under different indoor lighting conditions, avoiding reflections or dark areas that affect image quality.

[0023] 1.1.3 Image Processing Unit The image processing unit is integrated onto a circuit board inside the main body of the device. Its core is a single high-performance chip responsible for image processing tasks. This unit is divided into three functional modules via hardware circuitry, and these modules work collaboratively through an internal bus. Image preprocessing module: It uses the mean filtering algorithm to remove image noise, enhances image contrast through gamma correction, and performs grayscale processing to convert the color image into an 8-bit grayscale image, thereby improving the efficiency of subsequent processing. Feature extraction module: Extracts image edge features based on the Canny edge detection algorithm, analyzes scratch texture features by combining gray-level co-occurrence matrix, and collects color difference values ​​of scratch areas (compared with the vehicle body base color) and contour shape parameters (length, width, area, curvature) to form a multi-dimensional feature vector; Scratch recognition module: Built-in lightweight deep learning model (based on MobileNet architecture optimization). This model has been trained with 100,000+ vehicle scratch samples. It can receive feature vectors and match them with scratch templates and non-scratch texture (such as paint texture, stain) templates in the database. The recognition accuracy is ≥98%. It outputs the specific location of the scratch in the vehicle coordinate system, size (length accuracy ±0.1mm, width accuracy ±0.05mm), and shape type (linear, dotted, irregular).

[0024] 1.1.4 Data Storage and Analysis Unit This unit includes a 512GB SATA interface solid-state drive and a data analysis module: Solid-state drives are used to store raw acquired images, pre-processed images, and scratch recognition results. A partitioned storage strategy is adopted. Image files are named according to "acquisition time-vehicle identification code". Recognition results are stored in JSON format, which includes scratch ID, location coordinates, size parameters, recognition confidence, and other information. Data can be retained locally for more than 1 year. The data analysis module has built-in statistical algorithms that can generate statistical reports such as bar charts and pie charts by daily / weekly / monthly statistics on the number of scratches detected, the distribution of scratch types, and the location of high-frequency scratches, providing data support for maintenance plan development and insurance risk assessment.

[0025] 1.1.5 User Interface The user interface is integrated into the control panel of the main body of the device, including a capacitive touch screen and physical buttons; The touchscreen displays the device's operating status (standby / acquisition / processing / fault), real-time acquired images, scratch recognition and marking results, statistical reports, etc., and supports users to manually zoom in to view local images and export data files. Physical buttons are used for quick operation. The power button controls the start and stop of the device, the emergency stop button cuts off the device in abnormal situations, and the acquisition confirmation button supports manual triggering of image acquisition (for reshooting scenarios).

[0026] 1.1.6 Automation Control Unit This unit includes two ground sensors and one microcontroller; Ground sensors are installed at the entrance and exit of the equipment's detection area, embedded below the ground, and generate electrical signals by sensing the metal chassis of the vehicle. The microcontroller receives sensor signals and triggers the high-resolution camera to start acquiring images (continuously capturing 8-11 images covering the vehicle's detection surface) when the vehicle enters the entrance ground sensor area. When the vehicle exits the exit ground sensor area, the microcontroller controls the camera to stop acquiring images, ensuring the integrity and timeliness of image acquisition and realizing an automated detection process.

[0027] 1.1.7 Network Interface The device has a wired network interface (supporting Gigabit Ethernet) and a Wi-Fi module on the back. Users can establish communication with remote servers, maintenance management systems or insurance claims platforms through wired or wireless connections to transmit recognition results and image data in real time, supporting remote viewing and data sharing.

[0028] 1.2 Work Process Combination Figure 1 The workflow shown is illustrated below. The specific working steps of the device in this embodiment are as follows: Device startup: Press the power button and the device will perform a self-test (checking the status of components such as camera, sensor, and storage unit). After the self-test is passed, it will enter standby mode and the touch screen will display a "Ready" message. Triggered acquisition: The vehicle to be tested slowly drives into the designated detection area of ​​the equipment. After the entrance ground sensor detects the vehicle, it sends a trigger signal to the microcontroller. The microcontroller controls the LED lighting system to start (automatically adjusting the brightness according to the ambient light) and sends a photo-taking command to the high-resolution camera. The camera continuously acquires images of the vehicle surface (the acquisition angle covers the side or front of the vehicle, and is adjusted according to the equipment installation position). Image preprocessing: The acquired image is transmitted to the image processing unit through the internal data bus. The preprocessing module performs noise reduction, contrast enhancement and grayscale conversion in sequence to eliminate the impact of environmental noise and uneven lighting on image quality and output a clear preprocessed image. Feature extraction and scratch recognition: The feature extraction module extracts feature parameters such as edge, texture, color, and shape of scratches from the preprocessed image to form a feature vector; the scratch recognition module calls a deep learning model to analyze the feature vector, distinguish scratches from normal textures and stains on the vehicle surface, accurately identify scratch information, and generate recognition results including scratch location, size, and shape. Data storage and analysis: The recognition results, original images, and preprocessed images are synchronously stored on a solid-state drive. The data analysis module performs real-time statistics on the recognition results (such as the number of new scratches) and updates local statistical reports. Data display and transmission: The touchscreen displays the recognition results in real time (images of marked scratch areas + text descriptions), and users can view details through the touchscreen; at the same time, the device transmits the recognition results and image data to a remote server or related system through a network interface for subsequent maintenance assessment and insurance claims. Equipment Reset: After the vehicle leaves the exit ground sensor area, the microcontroller receives the signal, controls the LED lighting system to turn off, the camera to stop working, and the equipment returns to standby mode, waiting for the next detection task.

[0029] Example 2: This embodiment is based on Embodiment 1 and is optimized for outdoor detection scenarios (such as open-air parking lots and road rescue sites) to further improve the environmental adaptability and complex scratch recognition capabilities of the equipment, while still meeting the requirements of all claims of this invention.

[0030] 2.1 Optimize structural design 2.1.1 Image Acquisition Unit Optimization The high-resolution camera has been upgraded to a waterproof and dustproof type (IP67 protection rating), and the lens is equipped with a hydrophobic coating to prevent rain and dust from affecting image quality; a new lens sun hood has been added to reduce glare caused by direct sunlight. The LED lighting system now features a light-sensing module that can detect ambient light intensity in real time and automatically adjust LED brightness via a microcontroller to ensure image quality in all weather conditions.

[0031] 2.1.2 Image Processing Unit Optimization The embedded processor has been upgraded to a DSP chip, which improves the computing speed and supports parallel processing of complex scratches (such as cross scratches and mixed deep and shallow scratches). The scratch recognition module optimizes the deep learning model, adds a scratch depth grading function, and adds a depth parameter to the recognition results, providing more accurate data support for repair cost assessment.

[0032] 2.1.3 Equipment Main Body Optimization The outer shell is made of thickened aluminum alloy and has an added heat insulation layer, making it suitable for outdoor temperatures ranging from -40℃ to 85℃. The mounting bracket has been upgraded to a movable bracket with wheels (the wheels are equipped with a braking device), which allows for flexible deployment of the equipment in different outdoor locations. The bracket's height adjustment range has been expanded to accommodate large vehicles such as trucks and off-road vehicles.

[0033] 2.1.4 Data Security Optimization The data storage and analysis unit has added an AES-256 encryption module to encrypt the stored image data and recognition results to prevent data leakage. Network transmission uses the HTTPS protocol to ensure the security of remote data transmission.

[0034] 2.2 Work Process The working process of this embodiment is basically the same as that of embodiment 1, with the following optimization steps added: Ambient light detection: After the equipment is started, the light sensing module collects ambient light intensity data in real time and automatically adjusts the brightness parameters of the LED lighting system; Complex scratch recognition: During feature extraction, the image processing unit adds a branch recognition algorithm for intersecting scratches. It distinguishes the parameters of multiple intersecting scratches through contour segmentation technology and calculates the scratch depth through an optimized deep learning model. Encrypted Data Storage and Transmission: The recognition results and image data are encrypted and stored on a solid-state drive, and simultaneously transmitted to a remote server via HTTPS protocol. Data verification is performed during transmission to ensure data integrity. Mobile deployment: Users can push the device to the designated testing position using a movable bracket, lock the roller brake device, adjust the bracket height, and start the testing, adapting to flexible outdoor testing needs.

[0035] Example 3: This embodiment addresses the need for all-around scratch detection on the vehicle body. Based on embodiment 1, it increases the number of image acquisition units to achieve simultaneous detection in four directions: front, rear, left, and right. It is suitable for scenarios involving comprehensive scratch detection of the entire vehicle (such as vehicle annual inspection and used car appraisal).

[0036] 3.1 Multi-view structural design The main body of the equipment adopts a frame structure, with 4 sets of image acquisition units arranged around the detection area (corresponding to the front, rear, left and right of the vehicle body respectively). The structure of each set of acquisition units is the same as that of Example 1 (high resolution camera + LED cross lighting system). An automated control unit controls a set of ground sensors, and a microcontroller coordinates and controls four sets of image acquisition units to work synchronously (corresponding to the detection areas in four directions) to ensure that images of the vehicle body are acquired simultaneously from all directions, avoiding missed detections.

[0037] 3.2 Multi-view working process When the vehicle enters the frame-type detection area, the ground sensor is triggered, and four sets of image acquisition units start simultaneously to acquire images of the corresponding positions of the vehicle body. The image processing unit uses multi-threaded parallel processing technology to perform preprocessing, feature extraction and scratch recognition on four sets of images simultaneously. The recognition results are categorized, stored, and displayed according to the vehicle's orientation. The touchscreen supports split-screen viewing of scratch images from different orientations or synthesizing a full-vehicle scratch distribution diagram, allowing users to fully understand the vehicle's scratch situation. The data analysis module now includes a new function for tracking vehicle scratches, generating reports on the number of scratches, their distribution in all directions, and the percentage of scratches by severity, providing data support for used car valuation and comprehensive repair planning.

[0038] The above embodiments are merely preferred embodiments of the present invention and should not be construed as limiting the scope of protection of the present invention. Any non-substantial changes and substitutions made by those skilled in the art based on the present invention shall fall within the scope of protection claimed by the present invention.

Claims

1. A vehicle body scratch recognition and collection device, characterized in that, It includes an image acquisition unit, an image processing unit, a data storage and analysis unit, a user interface, and an automation control unit; The image acquisition unit is used to capture images of the vehicle body surface; the image processing unit is used to preprocess, extract features, and identify scratches on the acquired images; the data storage and analysis unit is used to store scratch data and perform statistical analysis; the user interface is used for device operation and data display; and the automation control unit is used to control the device operation and image acquisition process.

2. The vehicle body scratch recognition and acquisition device as described in claim 1, characterized in that, The image acquisition unit includes a high-resolution camera and an illumination device. The high-resolution camera has focus adjustment and exposure adjustment functions, and the illumination device is an LED illumination system that provides uniform and high-brightness light using a cross-illumination method.

3. The vehicle body scratch recognition and acquisition device as described in claim 1, characterized in that, The image processing unit includes an image preprocessing module, a feature extraction module, and a scratch recognition module. The image preprocessing module performs noise reduction, contrast enhancement, and grayscale conversion operations. The feature extraction module extracts the texture, color, and shape feature parameters of the scratches. The scratch recognition module uses a deep learning model to recognize and distinguish scratches.

4. The vehicle body scratch recognition and acquisition device as described in claim 1, characterized in that, The data storage and analysis unit includes a built-in large-capacity solid-state drive and a data analysis module. The solid-state drive stores the acquired images and recognition results, and the data analysis module performs statistical analysis on the location, size, and shape data of the scratches.

5. The vehicle body scratch recognition and acquisition device as described in claim 1, characterized in that, The user interface includes a touch screen display and physical buttons, providing an intuitive operating interface and data display method, and supporting touch operation and data viewing.

6. The vehicle body scratch recognition and acquisition device as described in claim 1, characterized in that, The automated control unit includes a photoelectric sensor or a ground sensor. When the vehicle enters the designated area, the sensor is triggered to generate a signal, which in turn controls the image acquisition unit to execute the photo-taking command, thereby realizing automated image acquisition.

7. The vehicle body scratch recognition and acquisition device as described in claim 1, characterized in that, The device also includes a network interface that supports Wi-Fi and wired network connections, and can transmit the identification results to a remote server or share data with other devices.

8. The vehicle body scratch recognition and acquisition device as described in claim 1, characterized in that, The device also includes a main body, the outer shell of which is made of aluminum alloy and has waterproof and dustproof functions. The main body is equipped with an adjustable height mounting bracket.

9. The vehicle body scratch recognition and acquisition device as described in claim 1, characterized in that, The image processing unit uses either a high-performance ARM or a deep learning chip as the main embedded processor chip, and is equipped with a DSP as an auxiliary chip to carry out image preprocessing, feature extraction and scratch recognition computational tasks.

Images