Vehicle damage assessment methods, devices, computer equipment and storage media
By integrating watermarking and parsing technology into vehicle damage assessment application software, and combining it with component segmentation models, the problems of high computational resource consumption and poor damage assessment results in existing vehicle damage assessment methods have been solved, achieving high accuracy in vehicle damage assessment.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- PING AN TECH (SHENZHEN) CO LTD
- Filing Date
- 2025-01-07
- Publication Date
- 2026-06-30
AI Technical Summary
Existing vehicle damage assessment methods suffer from problems such as high computational resource consumption and susceptibility to issues like non-standard image capture and visual angles when detecting damage to vehicle components, resulting in poor damage assessment results.
By integrating vehicle component detection models and damage detection models into vehicle damage assessment application software, and utilizing watermarking rules and parsing technology, combined with component segmentation models, damaged vehicle components can be accurately located, thereby improving the accuracy of damage assessment.
By analyzing the watermark information in vehicle images, the spatial and temporal context of vehicle damage can be accurately located, effectively improving the accuracy of vehicle damage assessment and ensuring the standardization and accuracy of the captured information.
Smart Images
Figure CN119810086B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of artificial intelligence technology, and in particular to a vehicle damage assessment method, apparatus, computer equipment, and storage medium. Background Technology
[0002] With the rapid development of technologies such as artificial intelligence, smartphones have become capable of integrating numerous lightweight models and applying them in various scenarios, providing users with a convenient, intelligent, and personalized experience. Currently, a relatively mature vehicle damage assessment method uses computer vision algorithms to detect vehicle components and damage, identifying damaged parts in images and recommending corresponding repair plans, achieving image-level vehicle damage assessment. However, this method requires deploying the vehicle component detection model and the vehicle damage detection model in a backend service cluster, resulting in significant computational resource consumption. Furthermore, when assessing vehicle damage from a single image, issues such as improper image capture and visual angles can lead to significant differences in the degree of damage to the same component between different images, further resulting in poor vehicle damage assessment results.
[0003] Therefore, improving the accuracy of vehicle damage assessment is a technical problem that urgently needs to be solved. Summary of the Invention
[0004] This invention provides a vehicle damage assessment method, apparatus, computer equipment, and storage medium to solve the technical problem that existing vehicle damage assessment methods are ineffective in assessing vehicle damage.
[0005] In a first aspect, a method for vehicle damage assessment application software is characterized by integrating a vehicle component detection model and a vehicle damage detection model within the software, the method comprising:
[0006] Based on the prompts from the vehicle damage assessment application software, the system intelligently takes photos of the vehicle currently awaiting damage assessment, thus obtaining the first image of the vehicle awaiting damage assessment.
[0007] A watermarking rule is preset for the first image of the vehicle to be damaged, and watermark information is added to the first image of the vehicle to be damaged according to the preset watermarking rule;
[0008] Obtain the second image of the vehicle to be assessed stored in the vehicle damage assessment application software, and determine whether the second image of the vehicle to be assessed contains watermark information;
[0009] The watermark information of the third vehicle image to be assessed is parsed from the second vehicle image to be assessed to obtain the shooting information of the third vehicle image to be assessed.
[0010] A component segmentation model for assessing damage to vehicle parts is preset, and the image of the third vehicle to be assessed and its capture information are input into the preset component segmentation model to identify the damage to the vehicle parts and obtain the assessment result of the vehicle parts.
[0011] Secondly, a vehicle damage assessment device is provided, the device being used to implement the vehicle damage assessment method as described in the first aspect above, comprising:
[0012] The shooting module is used to intelligently take pictures of the vehicle to be assessed based on the prompts from the vehicle damage assessment application software, and obtain the first image of the vehicle to be assessed.
[0013] The watermarking module is used to preset the watermarking rules for the first image of the vehicle to be damaged, and to add watermark information to the first image of the vehicle to be damaged according to the preset watermarking rules.
[0014] The judgment module is used to obtain the second image of the vehicle to be assessed stored in the vehicle damage assessment application software and to determine whether the second image of the vehicle to be assessed contains watermark information.
[0015] The watermark parsing module is used to parse the watermark information of the third vehicle image to be assessed from the second vehicle image to be assessed, so as to obtain the shooting information of the third vehicle image to be assessed.
[0016] The identification module is used to preset a component segmentation model for assessing the damage of vehicle components, and input the image of the third vehicle to be assessed and the image capture information of the third vehicle to be assessed into the preset component segmentation model to identify the damage to the vehicle components and obtain the damage assessment result of the vehicle components.
[0017] Thirdly, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described vehicle damage assessment method.
[0018] Fourthly, a computer-readable storage medium is provided, which stores a computer program that, when executed by a processor, implements the steps of the above-described vehicle damage assessment method.
[0019] In the above-described solution based on vehicle damage assessment methods, devices, computer equipment, and storage media, vehicle component detection models and vehicle damage detection models can be obtained through a client. These models are then integrated into a vehicle damage assessment software development kit (SDK), and the SSD is further integrated into a vehicle damage assessment application software. Based on prompts from the application software, the vehicle to be assessed is intelligently photographed to obtain a first image of the vehicle to be assessed. A watermarking rule is preset for the first image, and watermark information is added to it according to this rule. Finally, a second image of the vehicle to be assessed, stored in the application software, is obtained, and it is determined whether the second image contains water. The invention involves analyzing the watermark information in the second vehicle damage assessment software image to obtain the shooting information of the third vehicle damage assessment image. The third vehicle damage assessment image and its shooting information are then input into a component segmentation model for vehicle component damage identification, resulting in a damage assessment result for the vehicle components. In this invention, by determining whether the vehicle damage assessment image in the vehicle damage assessment software contains watermark information, the watermarked vehicle damage assessment image is analyzed. Through the traceable watermark information such as the shooting time and location of the analyzed vehicle damage assessment image, the spatiotemporal background of the vehicle damage can be accurately located, effectively helping damage assessors determine the specific damage condition of the vehicle at the time of damage, thereby effectively improving the accuracy of vehicle damage assessment. Attached Figure Description
[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0021] Figure 1 This is a schematic diagram of an application environment for a vehicle damage assessment method according to an embodiment of the present invention;
[0022] Figure 2 This is a schematic flowchart of a vehicle damage assessment method according to an embodiment of the present invention;
[0023] Figure 3 yes Figure 2 A schematic diagram of a specific implementation method for step S10;
[0024] Figure 4 yes Figure 2 A schematic diagram of a specific implementation method for step S20;
[0025] Figure 5yes Figure 2 A schematic diagram of a specific implementation method for step S30;
[0026] Figure 6 yes Figure 2 A schematic diagram of a specific implementation of step S40;
[0027] Figure 7 yes Figure 2 A schematic diagram of a specific implementation method for step S50;
[0028] Figure 8 This is a schematic diagram of a vehicle damage assessment device according to an embodiment of the present invention;
[0029] Figure 9 This is a schematic diagram of the structure of a computer device according to an embodiment of the present invention;
[0030] Figure 10 This is another structural schematic diagram of a computer device according to one embodiment of the present invention. Detailed Implementation
[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0032] The vehicle damage assessment method provided in this invention can be applied to, for example... Figure 1 In the application environment, Figure 1This is a schematic diagram of an application environment for a vehicle damage assessment method according to an embodiment of the present invention; wherein, the client communicates with the server via a network. The server can obtain a vehicle component detection model and a vehicle damage detection model from the client, integrate the obtained vehicle component detection model and vehicle damage detection model into a vehicle damage assessment software development kit, and integrate the vehicle damage assessment software development kit into a vehicle damage assessment application software; intelligently photograph the vehicle to be assessed according to the prompts of the vehicle damage assessment application software to obtain a first image of the vehicle to be assessed; preset watermarking rules for the first image of the vehicle to be assessed, and add watermark information to the first image of the vehicle to be assessed according to the preset watermarking rules; obtain a second image of the vehicle to be assessed stored in the vehicle damage assessment application software, and determine whether the second image of the vehicle to be assessed contains watermark information; process the second image of the vehicle to be assessed... The third image of the vehicle to be assessed, containing watermark information, is analyzed to obtain the shooting information of the third image. The third image and its shooting information are then input into a component segmentation model for damage identification of vehicle components, resulting in a damage assessment result. In this invention, by determining whether the images of the vehicles to be assessed in the vehicle damage assessment software contain watermark information, the watermarked images are analyzed. Through the traceable watermark information, such as the shooting time and location, the spatiotemporal context of the vehicle damage can be accurately located, effectively helping damage assessors determine the specific damage condition of the vehicle and thus significantly improving the accuracy of vehicle damage assessment. The invention will be described in detail below through specific embodiments.
[0033] Please see Figure 2 As shown, Figure 2 This is a schematic flowchart of a vehicle damage assessment method provided in an embodiment of the present invention. The vehicle damage assessment method specifically includes the following steps:
[0034] S10: Based on the prompts from the vehicle damage assessment application software, intelligently photograph the vehicle currently awaiting damage assessment to obtain the first image of the vehicle awaiting damage assessment.
[0035] This invention, when assessing vehicle damage, first integrates vehicle component detection models and damage detection models into a vehicle damage assessment software development kit (SDK). The SSD provides functions for calling and managing these models. Then, the SSD is integrated into the vehicle damage assessment application software, which customers can use to take standardized images. This ensures the comprehensiveness and accuracy of the vehicle information to be assessed, facilitating subsequent claims and repairs. For example, the vehicle damage assessment application software could be auto insurance claims software. In the auto insurance industry, insurance companies develop specialized claims software to simplify the vehicle damage assessment and claims process for customers. The integrated vehicle component detection model in this software can accurately identify key vehicle components such as license plates, vehicle identification numbers (VINs), headlights, doors, and bumpers. The integrated vehicle damage detection model can detect scratches, dents, collision damage, and other surface damage.
[0036] After integrating vehicle damage assessment software with vehicle component detection models and vehicle damage detection models, intelligent photography can be performed on the vehicle to be assessed based on the prompts from the vehicle damage assessment application software. This invention integrates a voice prompt module into the vehicle damage assessment software to guide the user in intelligently photographing the vehicle to be assessed. Specifically, for example… Figure 3 The above, Figure 3 yes Figure 2 A schematic flowchart of a specific implementation of step S10 includes the following steps S11-S14:
[0037] S11: Based on the vehicle component detection model integrated in the vehicle damage assessment application software, the video frames in the real-time image are analyzed to detect whether the real-time image contains the license plate number and vehicle identification number (VIN) of the vehicle to be assessed. Specifically, this invention can first use the camera function of a mobile device to take pictures of the vehicle to be assessed; then, based on the vehicle component detection model integrated in the vehicle damage assessment application software, the video frames in the real-time image are analyzed to detect whether the real-time image contains the license plate number and VIN of the vehicle to be assessed.
[0038] S12: If the real-time image contains the license plate number and vehicle identification number (VIN) of the vehicle to be assessed, then a distant view of the damaged area of the vehicle to be assessed is taken according to the voice prompts of the vehicle damage assessment application software. Specifically, if the real-time image contains the license plate number and VIN of the vehicle to be assessed in step S11, the vehicle damage assessment application software will prompt the user via voice, "The license plate number and VIN have been recognized. Please take a distant view of the damaged area."
[0039] S13: Based on the vehicle damage detection model integrated in the vehicle damage assessment application software, the distant view of the part to be inspected of the vehicle to be assessed is analyzed to detect whether the part to be inspected of the vehicle to be assessed contains a damaged area. Specifically, after taking a distant view of the part to be inspected of the vehicle to be assessed, this invention needs to continue to use the damage detection model integrated in the vehicle damage assessment application software to perform a preliminary analysis of the distant view taken by the user to confirm whether the image contains a possible damaged area.
[0040] S14: If a damaged area is detected in the part of the vehicle to be assessed, a close-up image of the part of the vehicle to be assessed is taken according to the voice prompts of the vehicle damage assessment application software to obtain a first image of the vehicle to be assessed. Specifically, if a suspected damaged area of the vehicle to be assessed is detected in step S13, the user is prompted by voice to "Please take a close-up image of the damaged area".
[0041] Furthermore, during the intelligent photography process, the software development kit (SDK) within the vehicle damage assessment application will provide real-time feedback on the detection results to the user. For example, if the license plate or vehicle identification number (VIN) is unclear, the software will provide a voice prompt stating, "The license plate / VIN is unclear; please retake the photo." For both distant and close-up views of the vehicle, if the shooting angle, distance, or other factors do not meet the requirements, the software will provide voice prompts to guide the user to retake the photo to obtain a standardized first image of the vehicle to be assessed.
[0042] S20: Preset watermarking rules for the first image of the vehicle to be assessed, and add watermark information to the image according to the preset rules. Specifically, in actual use scenarios, customers often need to forward the first image of the vehicle to be assessed via WeChat or other software. To enable the storage and retrieval of information from the shooting stage, we added watermarking to the processing stage of the vehicle damage assessment application software to ensure effective parsing of the shooting information from the first image of the vehicle to be assessed. Specifically, such as... Figure 4 The above, Figure 4 yes Figure 2 A schematic flowchart of a specific implementation of step S20 includes the following steps S21-S22:
[0043] S21: Based on the size information of the first image of the vehicle to be damaged, add a white border of appropriate size to the bottom of the first image of the vehicle to be damaged, and update the size information of the first image of the vehicle to be damaged. Specifically, this invention first defines the rules for adding a watermark to the first image of the vehicle to be damaged, that is, adding a white border to the bottom of the first image of the vehicle to be damaged according to a corresponding ratio, so that the ratio of the width to the height of the white border is consistent with the ratio of the width to the height of the original first image of the vehicle to be damaged, and then updating the size information of the first image of the vehicle to be damaged after adding the white border information.
[0044] S22: Add watermark information to the first image of the vehicle to be assessed after updating its size information according to a preset text content format, to obtain the first image of the vehicle to be assessed with watermark information. Specifically, after adding a white border to the first image of the vehicle to be assessed, the present invention adds text content to the far right of the white border. The text content is in the format of "whether it is a smart shot + component code + whether it is a detail image", to obtain the first image of the vehicle to be assessed with watermark information.
[0045] S30: Obtain the second image of the vehicle to be assessed stored in the vehicle damage assessment application software, and determine whether the second image contains watermark information. Before parsing the watermark from the second image of the vehicle to be assessed stored in the vehicle damage assessment application software, it is necessary to determine whether the second image contains watermark information. Specifically, such as... Figure 5 The above, Figure 5 yes Figure 2 A schematic flowchart of a specific implementation of step S30 includes the following steps S41-S44:
[0046] S31: Convert the second image of the vehicle to be damaged to RGB format to obtain the fourth image of the vehicle to be damaged. Specifically, to determine whether the image of the vehicle to be damaged contains watermark information, the present invention first converts the second image of the vehicle to be damaged to RGB format to obtain the fourth image of the vehicle to be damaged after RGB format conversion. Since computer monitors display images based on the RGB color model, if the second image of the vehicle to be damaged is not in RGB format, the computer monitor needs to perform RGB format conversion before displaying it to facilitate the processing of each color channel in the image of the vehicle to be damaged.
[0047] S32: The height ratio of the white border in the fourth image of the vehicle to be damaged is preset, and the height range of the white border in the fourth image of the vehicle to be damaged is predicted based on the size information of the fourth image of the vehicle to be damaged and the preset height ratio of the white border in the fourth image of the vehicle to be damaged. Specifically, this invention can calculate the theoretical height value of the white border based on the preset height ratio of the white border in the fourth image of the vehicle to be damaged, thereby determining the height range that may contain a watermark. For example, if the preset height ratio of the white border in the fourth image of the vehicle to be damaged is 1 / 10 of the original image height, then the height range of the white border in the bottom area where the watermark is located can be predicted based on the total height of the fourth image of the vehicle to be damaged.
[0048] S33: Based on the predicted white border height range of the fourth image of the vehicle to be damaged, the white border area of the fourth image of the vehicle to be damaged is selected through a center cropping algorithm. Specifically, in different operating systems, due to factors such as image rendering mechanisms and differences in device resolution, pixel-level errors may occur when adding watermarks. This error may cause the actual position and size of the white border to deviate slightly from the theoretically calculated value. Therefore, it is necessary to further determine the range of the white border. The fourth image of the vehicle to be damaged can be center-cropped to select the specific area of the white border.
[0049] S34: Determine whether the white-edge region of the fourth vehicle image to be assessed contains only white pixels based on the image pixel value analysis function. Specifically, this invention uses the image pixel value analysis function to determine whether the white-edge region of the selected fourth vehicle image to be assessed contains only pixels with values of (255, 255, 255). If the result returned by the image pixel value analysis function indicates that the region contains only pixels with a value of 255 in both the RGB channels, then the region can be determined to be a white-edge region.
[0050] S35: If the white border area of the fourth image of the vehicle to be assessed contains only white pixel values, then it is determined that the second image of the vehicle to be assessed contains watermark information. Specifically, if the white border area of the fourth image of the vehicle to be assessed is white, then the right half of the white border area obtained in step S42 needs to be saved. If the white border area of the fourth image of the vehicle to be assessed is not white, then it means that the image does not contain the watermark we defined, and no further watermark parsing operation is required.
[0051] S40: Perform watermark parsing on the third image of the vehicle to be assessed, which contains watermark information in the second image of the vehicle to be assessed, to obtain the capturing information of the third image of the vehicle to be assessed. If it is detected that the second image of the vehicle to be assessed stored in the vehicle damage assessment application software contains watermark information, then it is necessary to parse the third image of the vehicle to be assessed containing the watermark information. Specifically, such as... Figure 6 The above, Figure 6 yes Figure 2A schematic flowchart of a specific implementation of step S40 includes the following steps S41-S44:
[0052] S41: The text detection module in the optical character recognition algorithm is used to perform text detection on the image of the third vehicle to be damaged, thereby obtaining the text region of the image. Specifically, this invention uses the text detection module to extract features from the image of the third vehicle to be damaged. By extracting feature information such as color changes, edge information, and texture features from the image, the text region of the image is determined.
[0053] S42: The text recognition module in the optical character recognition algorithm performs character recognition on the detected text region of the third vehicle to be assessed for damage, thereby obtaining the character information of the image of the third vehicle to be assessed for damage. Specifically, this invention uses the text recognition module in the optical character recognition algorithm to match the extracted text region features with the standard features of various characters stored in the model. By calculating similarity, probability, and other indicators, the most likely corresponding character in each text region is determined, thereby converting the image content in the text region into specific character information, and finally obtaining the character information contained in the image of the third vehicle to be assessed for damage.
[0054] S43: Preset watermarking rules for the third image of the vehicle to be damaged, and use regular expressions to determine whether the character information of the third image meets the preset watermarking rules. The watermarking rules for the third image of the vehicle to be damaged are set according to actual needs and business logic. For example, the preset watermarking rules for the third image of the vehicle to be damaged can be: "Whether it was intelligently captured + component code + whether it is a detailed image". Based on the preset watermarking rules for the third image of the vehicle to be damaged, regular expressions can be used to determine whether the character information of the third image meets the preset watermarking rules.
[0055] S44: If the character information of the third vehicle image to be assessed meets the preset watermarking rules, the shooting information of the third vehicle image to be assessed is extracted. When it is determined that the character information of the third vehicle image to be assessed meets the preset watermarking rules, the shooting information of the third vehicle image to be assessed is extracted based on the character information. For example, the shooting date, shooting location, and the serial number of the equipment used for shooting are extracted. This shooting information can be used for subsequent vehicle damage assessment records, data archiving, traceability, and other related work.
[0056] S50: A component segmentation model for assessing damage to vehicle parts is preset, and the image of the third vehicle to be assessed and its capture information are input into the preset component segmentation model to identify the damage to the vehicle parts, thereby obtaining the damage assessment result of the vehicle parts. Specifically, as follows... Figure 7The above, Figure 7 yes Figure 2 A schematic flowchart of a specific implementation of step S50 includes the following steps S51-S53:
[0057] S51: Extract the feature information of the third image of the vehicle to be damaged, and fuse the extracted feature information of the third image of the vehicle to be damaged with the shooting information of the third image of the vehicle to be damaged to obtain the target feature information of the third image of the vehicle to be damaged. Specifically, the shooting information of the third image of the vehicle to be damaged can be used as a priori condition to assist the component segmentation model in identifying damaged vehicle components, thereby improving the component segmentation model's ability to identify the third image of the vehicle to be damaged. Combining the captured component information with the component segmentation model can more accurately locate the component area and reduce the possibility of false detection of components due to differences in vehicle models, image quality, etc. For example, if the captured component is a headlight, then the image should be near the headlight, not the rear of the vehicle. Combining the detail image with the component segmentation model results can reduce the number of missed detections of components. If the captured image is a detail of the left front fender, the component segmentation may only be able to determine that it is a fender and cannot determine the location due to the excessive detail. In this case, the location information captured can be used for subsequent judgment.
[0058] S52: Input the target feature information of the fused third image of the vehicle to be damaged into a preset component segmentation model to divide the components of the vehicle to be damaged into regions. Specifically, this invention inputs the target feature information of the fused third image of the vehicle to be damaged into the component segmentation model. The component segmentation model outputs a segmentation map with the same size as the input image. By parsing the segmentation map, the regions where each component of the vehicle to be damaged is located can be divided.
[0059] S53: Determine the degree of damage to each area of the divided vehicle component to be assessed, and obtain the damage assessment result for the vehicle component. Specifically, the degree of damage to each area of the divided vehicle component can be determined by observing the changes in the appearance characteristics of each area, such as whether there are scratches, dents, etc. on the surface of the component.
[0060] In one embodiment of the present invention, the vehicle damage assessment method further includes: preprocessing the data of the third image of the vehicle to be assessed, and fusing the preprocessed data of the third image of the vehicle to be assessed into the training set of the vehicle component detection model and the vehicle damage detection model. Specifically, the present invention can also backflow the parsed data of the third image of the vehicle to be assessed, iterate the model optimization algorithm, and since detailed images of the vehicle to be assessed can capture minor damage, they often cannot determine the location information of components. Therefore, the data of the third image of the vehicle to be assessed can be expanded into the training set of the vehicle component detection model and the vehicle damage detection model to improve the vehicle damage assessment software's ability to identify and generalize close-up images and minor damage. In addition, the fusion of multimodal information is also considered, taking into account multi-source information such as other sensor data and vehicle history records, and making comprehensive judgments. The complementary advantages of these multimodal information are utilized to improve the comprehensive understanding and judgment ability of vehicle damage status.
[0061] As can be seen, in the above solution, when assessing vehicle damage, the accuracy of vehicle damage assessment is effectively improved by determining whether the vehicle images to be assessed in the vehicle damage assessment software contain watermark information, analyzing the watermarked vehicle images, and identifying the damage to vehicle parts based on the analyzed shooting information.
[0062] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.
[0063] In one embodiment, a vehicle damage assessment device is provided, which corresponds one-to-one with the vehicle damage assessment methods described in the above embodiments. For example... Figure 8 As shown, Figure 8 This is a schematic diagram of a vehicle damage assessment device according to an embodiment of the present invention. The vehicle damage assessment device includes a shooting module 81, a watermark adding module 82, a judgment module 83, a watermark parsing module 84, and an identification module 85. Detailed descriptions of each functional module are as follows:
[0064] The shooting module 81 is used to intelligently take pictures of the vehicle to be assessed based on the prompts from the vehicle damage assessment application software, and obtain the first image of the vehicle to be assessed.
[0065] The watermarking module 82 is used to preset the watermarking rules for the first image of the vehicle to be damaged, and to add watermark information to the first image of the vehicle to be damaged according to the preset watermarking rules.
[0066] The judgment module 83 is used to obtain the second image of the vehicle to be assessed stored in the vehicle damage assessment application software, and to determine whether the second image of the vehicle to be assessed contains watermark information.
[0067] The watermark parsing module 84 is used to perform watermark parsing on the third image of the vehicle to be damaged that contains watermark information in the second image of the vehicle to be damaged, so as to obtain the shooting information of the third image of the vehicle to be damaged.
[0068] The identification module 85 is used to preset a component segmentation model for assessing the damage of vehicle components, and input the image of the third vehicle to be assessed and the image capture information of the third vehicle to be assessed into the preset component segmentation model to identify the damage of the vehicle components and obtain the damage assessment result of the vehicle components.
[0069] In one embodiment, the imaging module 81 is specifically used for:
[0070] Based on the vehicle component detection model integrated in the vehicle damage assessment application software, the video frames in the real-time image are analyzed to detect whether the real-time image contains the license plate number and vehicle identification number of the vehicle to be assessed.
[0071] If the license plate and vehicle identification number of the vehicle to be assessed are detected in the real-time image, a distant view of the part of the vehicle to be assessed is taken according to the voice prompts of the vehicle damage assessment application software.
[0072] Based on the vehicle damage detection model integrated in the vehicle damage assessment application software, the remote view of the part to be inspected of the vehicle to be assessed is analyzed to detect whether the part to be inspected of the vehicle to be assessed contains a damaged area.
[0073] If the detected part of the vehicle to be assessed contains a damaged area, a close-up image of the detected part of the vehicle to be assessed is taken according to the voice prompts of the vehicle damage assessment application software to obtain a first image of the vehicle to be assessed.
[0074] In one embodiment, the watermark adding module 82 is specifically used for:
[0075] Based on the size information of the first image of the vehicle to be damaged, a white border of the corresponding size is added to the bottom of the first image of the vehicle to be damaged, and the size information of the first image of the vehicle to be damaged is updated.
[0076] Watermark information is added to the first image of the vehicle to be assessed after updating the size information according to the preset text content format, resulting in the first image of the vehicle to be assessed with watermark information.
[0077] In one embodiment, the determination module 83 is specifically used for:
[0078] The second image of the vehicle to be assessed for damage is converted to RGB format to obtain the fourth image of the vehicle to be assessed for damage.
[0079] The height ratio of the white border in the fourth image of the vehicle to be damaged is preset, and the height range of the white border in the fourth image of the vehicle to be damaged is predicted based on the size information of the fourth image of the vehicle to be damaged and the preset height ratio of the white border in the fourth image of the vehicle to be damaged.
[0080] Based on the predicted white border height range of the fourth image of the vehicle to be damaged, the white border area of the fourth image of the vehicle to be damaged is selected by the center cropping algorithm.
[0081] Based on the image pixel value analysis function, determine whether the white border area of the fourth vehicle to be damaged contains only white pixel values;
[0082] If the white border area of the fourth image of the vehicle to be assessed contains only white pixel values, then it is determined that the second image of the vehicle to be assessed contains watermark information.
[0083] In one embodiment, the watermark parsing module 84 is specifically used for:
[0084] The text detection module in the optical character recognition algorithm is used to perform text detection on the image of the third vehicle to be damaged, and the text region of the image of the third vehicle to be damaged is obtained.
[0085] The text recognition module in the optical character recognition algorithm performs character recognition on the text region of the detected third vehicle to be damaged, and obtains the character information of the image of the third vehicle to be damaged.
[0086] A watermarking rule is preset for the third image of the vehicle to be assessed for damage, and a regular expression is used to determine whether the character information of the third image of the vehicle to be assessed for damage meets the preset watermarking rule.
[0087] If the character information of the third image of the vehicle to be damaged meets the preset watermarking rules, the shooting information of the third image of the vehicle to be damaged is extracted.
[0088] In one embodiment, the identification module 85 is specifically used for:
[0089] Extract the feature information of the third image of the vehicle to be damaged, and fuse the extracted feature information of the third image of the vehicle to be damaged with the shooting information of the third image of the vehicle to be damaged to obtain the target feature information of the third image of the vehicle to be damaged.
[0090] The target feature information of the fused third image of the vehicle to be damaged is input into the preset component segmentation model to divide the components of the vehicle to be damaged into regions.
[0091] The degree of damage to each area of the pre-defined vehicle component to be assessed is determined to obtain the damage assessment result for the vehicle component.
[0092] In one embodiment, the vehicle damage assessment device is further configured to:
[0093] The data of the third vehicle to be damaged is preprocessed, and the preprocessed data of the third vehicle to be damaged is fused into the training set of the vehicle component detection model and the vehicle damage detection model.
[0094] This invention provides a vehicle damage assessment device. When assessing vehicle damage, it determines whether the vehicle image to be assessed in the vehicle damage assessment software contains watermark information. The device analyzes the watermarked vehicle image and uses traceable watermark information such as the shooting time and location of the image to accurately locate the spatiotemporal context of the vehicle damage. This effectively helps damage assessors determine the specific damage condition of the vehicle at the time of the damage, thereby significantly improving the accuracy of vehicle damage assessment.
[0095] For specific limitations regarding the vehicle damage assessment device, please refer to the limitations on the vehicle damage assessment method above, which will not be repeated here. Each module in the aforementioned vehicle damage assessment device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.
[0096] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 9 As shown, Figure 9 This is a schematic diagram of a computer device according to an embodiment of the present invention. The computer device includes a processor, a memory, a network interface, and a database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface of the computer device is used to communicate with external clients via a network connection. When the computer program is executed by the processor, it implements the functions or steps of a vehicle damage assessment method on the server side.
[0097] In one embodiment, a computer device is provided, which may be a client, and its internal structure diagram may be as follows: Figure 10 As shown, Figure 10This is another schematic diagram of a computer device according to an embodiment of the present invention. The computer device includes a processor, memory, network interface, display screen, and input device connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface is used to communicate with an external server via a network connection. When the computer program is executed by the processor, it implements the functions or steps of a vehicle damage assessment method on the client side.
[0098] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps:
[0099] Based on the prompts from the vehicle damage assessment application software, the system intelligently takes photos of the vehicle currently awaiting damage assessment, thus obtaining the first image of the vehicle awaiting damage assessment.
[0100] A watermarking rule is preset for the first image of the vehicle to be damaged, and watermark information is added to the first image of the vehicle to be damaged according to the preset watermarking rule;
[0101] Obtain the second image of the vehicle to be assessed stored in the vehicle damage assessment application software, and determine whether the second image of the vehicle to be assessed contains watermark information;
[0102] The watermark information of the third vehicle image to be assessed is parsed from the second vehicle image to be assessed to obtain the shooting information of the third vehicle image to be assessed.
[0103] A component segmentation model for assessing damage to vehicle parts is preset, and the image of the third vehicle to be assessed and its capture information are input into the preset component segmentation model to identify the damage to the vehicle parts and obtain the assessment result of the vehicle parts.
[0104] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:
[0105] Based on the prompts from the vehicle damage assessment application software, the system intelligently takes photos of the vehicle currently awaiting damage assessment, thus obtaining the first image of the vehicle awaiting damage assessment.
[0106] A watermarking rule is preset for the first image of the vehicle to be damaged, and watermark information is added to the first image of the vehicle to be damaged according to the preset watermarking rule;
[0107] Obtain the second image of the vehicle to be assessed stored in the vehicle damage assessment application software, and determine whether the second image of the vehicle to be assessed contains watermark information;
[0108] The watermark information of the third vehicle image to be assessed is parsed from the second vehicle image to be assessed to obtain the shooting information of the third vehicle image to be assessed.
[0109] A component segmentation model for assessing damage to vehicle parts is preset, and the image of the third vehicle to be assessed and its capture information are input into the preset component segmentation model to identify the damage to the vehicle parts and obtain the assessment result of the vehicle parts.
[0110] It should be noted that the functions or steps that can be implemented by the computer-readable storage medium or computer device described above can be referred to the relevant descriptions on the server side and client side in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.
[0111] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.
[0112] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.
[0113] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.
Claims
1. A vehicle damage assessment method, applied to vehicle damage assessment application software, characterized in that, The vehicle damage assessment application software integrates vehicle component detection models and vehicle damage detection models, and the method includes: The vehicle component detection model and the vehicle damage detection model are integrated into the vehicle damage assessment software development kit, and the vehicle damage assessment software development kit is integrated into the vehicle damage assessment application software; Based on the prompts from the vehicle damage assessment application software, intelligent photography is performed on the vehicle currently awaiting damage assessment to obtain the first image of the vehicle awaiting assessment, including: Based on the vehicle component detection model integrated in the vehicle damage assessment application software, the video frames in the real-time image are analyzed to detect whether the real-time image contains the license plate number and vehicle identification number of the vehicle to be assessed. If the license plate and vehicle identification number of the vehicle to be assessed are detected in the real-time image, a distant view of the part of the vehicle to be assessed is taken according to the voice prompts of the vehicle damage assessment application software. Based on the vehicle damage detection model integrated in the vehicle damage assessment application software, the remote view of the part to be inspected of the vehicle to be assessed is analyzed to detect whether the part to be inspected of the vehicle to be assessed contains a damaged area. If the detected part of the vehicle to be assessed contains a damaged area, a close-up image of the detected part of the vehicle to be assessed is taken according to the voice prompt information of the vehicle damage assessment application software to obtain the first image of the vehicle to be assessed. A watermarking rule is preset for the first image of the vehicle to be damaged, and watermark information is added to the first image of the vehicle to be damaged according to the preset watermarking rule; Obtain the second image of the vehicle to be assessed stored in the vehicle damage assessment application software, and determine whether the second image of the vehicle to be assessed contains watermark information; The watermark information of the third vehicle image to be assessed is parsed from the second vehicle image to be assessed to obtain the shooting information of the third vehicle image to be assessed. A component segmentation model for assessing damage to vehicle parts is preset, and the image of the third vehicle to be assessed and its capture information are input into the preset component segmentation model to identify the damage to the vehicle parts and obtain the assessment result of the vehicle parts.
2. The vehicle damage assessment method according to claim 1, characterized in that, The preset watermarking rules for the first image of the vehicle to be damaged, and the addition of watermark information to the image of the first image of the vehicle to be damaged according to the preset watermarking rules, include: Based on the size information of the first image of the vehicle to be damaged, a white border of the corresponding size is added to the bottom of the first image of the vehicle to be damaged, and the size information of the first image of the vehicle to be damaged is updated. Watermark information is added to the first image of the vehicle to be assessed after updating the size information according to the preset text content format, resulting in the first image of the vehicle to be assessed with watermark information.
3. The vehicle damage assessment method according to claim 1, characterized in that, The determination of whether the second image of the vehicle to be damaged contains watermark information includes: The second image of the vehicle to be assessed for damage is converted to RGB format to obtain the fourth image of the vehicle to be assessed for damage. The height ratio of the white border in the fourth image of the vehicle to be damaged is preset, and the height range of the white border in the fourth image of the vehicle to be damaged is predicted based on the size information of the fourth image of the vehicle to be damaged and the preset height ratio of the white border in the fourth image of the vehicle to be damaged. Based on the predicted white border height range of the fourth image of the vehicle to be damaged, the white border area of the fourth image of the vehicle to be damaged is selected by the center cropping algorithm. Based on the image pixel value analysis function, determine whether the white border area of the fourth vehicle to be damaged contains only white pixel values; If the white border area of the fourth image of the vehicle to be assessed contains only white pixel values, then it is determined that the second image of the vehicle to be assessed contains watermark information.
4. The vehicle damage assessment method according to claim 1, characterized in that, The process involves parsing the watermark information from the second image of the vehicle to be assessed, which contains watermark information, to obtain the capturing information of the third image of the vehicle to be assessed, including: The text detection module in the optical character recognition algorithm is used to perform text detection on the image of the third vehicle to be damaged, and the text region of the image of the third vehicle to be damaged is obtained. The text recognition module in the optical character recognition algorithm performs character recognition on the text region of the detected third vehicle to be damaged, and obtains the character information of the image of the third vehicle to be damaged. A watermarking rule is preset for the third image of the vehicle to be assessed for damage, and a regular expression is used to determine whether the character information of the third image of the vehicle to be assessed for damage meets the preset watermarking rule. If the character information of the third image of the vehicle to be damaged meets the preset watermarking rules, the shooting information of the third image of the vehicle to be damaged is extracted.
5. The vehicle damage assessment method according to claim 1, characterized in that, The step of inputting the image of the third vehicle to be damaged and the image capture information into a preset component segmentation model to identify the damage to vehicle components and obtain the damage assessment result of the vehicle components includes: Extract the feature information of the third image of the vehicle to be damaged, and fuse the extracted feature information of the third image of the vehicle to be damaged with the shooting information of the third image of the vehicle to be damaged to obtain the target feature information of the third image of the vehicle to be damaged. The target feature information of the fused third image of the vehicle to be damaged is input into the preset component segmentation model to divide the components of the vehicle to be damaged into regions. The degree of damage to each area of the pre-defined vehicle component to be assessed is determined to obtain the damage assessment result for the vehicle component.
6. The vehicle damage assessment method according to claim 1, characterized in that, The vehicle damage assessment method also includes: The data of the third vehicle to be damaged is preprocessed, and the preprocessed data of the third vehicle to be damaged is fused into the training set of the vehicle component detection model and the vehicle damage detection model.
7. A vehicle damage assessment device, said device being used to implement the vehicle damage assessment method as described in any one of claims 1-6, characterized in that, include: The shooting module is used to intelligently take pictures of the vehicle to be assessed based on the prompts from the vehicle damage assessment application software, and obtain the first image of the vehicle to be assessed. The watermarking module is used to preset the watermarking rules for the first image of the vehicle to be damaged, and to add watermark information to the first image of the vehicle to be damaged according to the preset watermarking rules. The judgment module is used to obtain the second image of the vehicle to be assessed stored in the vehicle damage assessment application software and to determine whether the second image of the vehicle to be assessed contains watermark information. The watermark parsing module is used to parse the watermark information of the third vehicle image to be assessed from the second vehicle image to be assessed, so as to obtain the shooting information of the third vehicle image to be assessed. The identification module is used to preset a component segmentation model for assessing the damage of vehicle components, and input the image of the third vehicle to be assessed and the image capture information of the third vehicle to be assessed into the preset component segmentation model to identify the damage to the vehicle components and obtain the damage assessment result of the vehicle components.
8. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the vehicle damage assessment method as described in any one of claims 1 to 6.
9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the vehicle damage assessment method as described in any one of claims 1 to 6.