Vehicle insurance survey photographing guiding method and system based on multi-model fusion

Abstract

The invention discloses a vehicle insurance survey photographing guiding method and system based on multi-model fusion, and the method comprises the steps: carrying out the detection of visual data containing a target vehicle based on a vehicle detection model, and obtaining the position information; identifying the position information by using a vehicle orientation identification model to obtain an orientation and a shooting distance; identifying the visual data through an information identification algorithm to obtain a license plate and a VIN code; determining a corresponding guide rule according to the vehicle insurance case reporting information; outputting photographing guide information based on the position information, the orientation, the photographing distance, the license plate, the VIN code and the guide rule; the photographing progress is updated in real time through a visual interface of the mobile terminal, and the uncompleted photographing orientation and the completed photographing orientation are displayed in a distinguished mode through color changes in the visual interface; and carrying out associated storage on an image shot according to the shooting guide information and the license plate and / or the VIN code. The photographing efficiency and the picture quality can be improved, the labor cost is reduced, and the vehicle insurance claim settlement period is shortened.

Description

Technical Field

[0001] This invention relates to the fields of computer vision and mobile application technology, and in particular to a method and system for guiding vehicle insurance inspection by taking photos based on multi-model fusion. Background Technology

[0002] Taking photos during vehicle insurance claims inspections is a crucial step in the mobile claims process. Traditional mobile claims inspection photography relies entirely on manual operation, lacking intelligent technology support. Inspectors need to use their personal experience to determine the shooting location, such as the four corners of the vehicle, or its front and rear positions; determine the shooting distance, such as long shot, medium shot, or close-up; and control the number of photos required. Related applications only provide basic photo taking and storage functions, lacking vehicle detection capabilities, the ability to automatically identify the presence and specific location of vehicles in images, and any intelligent guidance mechanisms, including location prompts, quantity verification, and distance suggestions.

[0003] Currently, existing technologies have three main shortcomings: First, the lack of vehicle detection capabilities makes it difficult to locate the target in photos. Because it is impossible to identify whether a vehicle exists in the image and its specific location, surveyors are prone to taking photos of non-vehicle scenes or missing key parts of the vehicle, resulting in invalid photos. This may require re-photographing, increasing labor costs and the claims processing time.

[0004] Secondly, the lack of intelligent guidance mechanisms results in poor operational standardization. Existing technology fails to provide directional guidance, such as clearly defining standard shooting angles like 45 degrees to the left front or right rear; it lacks quantity verification, for example, failing to remind users that at least one photo is required for each of the four corners and at least three close-up photos of the vehicle damage; and it doesn't offer distance suggestions, such as not distinguishing between long-range, medium-range, and close-up shots. The photo-taking process relies entirely on the surveyor's experience, leading to significant differences in shooting standards among different personnel and inconsistent photo quality, which affects the subsequent assessment of vehicle damage in the claims process.

[0005] Third, the lack of data association capabilities makes it difficult to support subsequent claims processes. Current technology lacks automatic recognition and storage functions for license plates and VIN codes, requiring manual entry of vehicle identification information, which is prone to errors and inefficient. Furthermore, photos cannot be automatically linked to key information such as vehicle identity, shooting location, and shooting distance, making it impossible to quickly locate the required photos in subsequent stages, necessitating a review of each photo and impacting claims efficiency.

[0006] Therefore, there is an urgent need to provide a technical solution to address the above problems. Summary of the Invention

[0007] To address the aforementioned technical problems, this invention provides a method, system, electronic device, and storage medium for guiding vehicle insurance survey photography based on multi-model fusion.

[0008] Firstly, this invention provides a vehicle insurance survey photo guidance method based on multi-model fusion, the technical solution of which is as follows: Based on the vehicle detection model, visual data containing the target vehicle collected through a mobile terminal is detected to obtain the location information of the target vehicle; The location information is identified using a vehicle orientation recognition model to obtain the orientation and shooting distance of the target vehicle; The visual data is identified using an information recognition algorithm to obtain the license plate and VIN code of the target vehicle. Determine the accident type based on the car insurance claim information and obtain the corresponding guidance rules for the accident type; Based on the location information, orientation, shooting distance, license plate, VIN code, and guidance rules, photo-taking guidance information is output; wherein, the photo-taking guidance information includes: orientation prompts, quantity verification, and distance suggestions; During the process of outputting the photo-taking guidance information, the photo-taking progress is updated in real time through the visual interface of the mobile terminal, and the incomplete and completed shooting locations are distinguished by color changes in the visual interface. The image taken according to the photo-taking guidance information will be associated with and stored with the license plate and / or the VIN code.

[0009] The beneficial effects of the vehicle insurance survey photo guidance method based on multi-model fusion of the present invention are as follows: The method of this invention can solve the problems of difficult target positioning, non-standard operation and weak data correlation caused by the lack of photo guidance in traditional auto insurance surveys, improve photo efficiency and photo quality, reduce labor costs and shorten the auto insurance claims cycle.

[0010] Based on the above scheme, the vehicle insurance survey photo guidance method based on multi-model fusion of the present invention can be further improved as follows.

[0011] In one optional approach, the vehicle detection model includes: a feature extraction network, a feature pyramid network, and a detection head; the step of detecting visual data containing the target vehicle collected via a mobile terminal based on the vehicle detection model to obtain the location information of the target vehicle includes: The mobile terminal acquires the visual data containing the target vehicle in real time. The visual data is input into the feature extraction network of the vehicle detection model to extract multi-scale feature maps of the visual data. The multi-scale feature map is input into the feature pyramid network for feature fusion to generate enhanced multi-scale features; The enhanced multi-scale features are processed by the detection head to regress and generate the bounding box coordinates of the target vehicle. The position information of the target vehicle in the visual data is determined based on the bounding box coordinates.

[0012] The beneficial effects of adopting the above-mentioned optional approach are as follows: by further enhancing the adaptability of the vehicle detection model to complex scenes and different shooting conditions through the multi-scale feature fusion architecture of feature extraction network, feature pyramid network and detection head, the detection accuracy and robustness of target vehicle location information are significantly improved.

[0013] In one optional embodiment, the vehicle orientation recognition model includes: a plurality of convolutional blocks connected in sequence and a fully connected layer; the step of using the vehicle orientation recognition model to identify the location information and obtain the orientation and shooting distance of the target vehicle includes: Based on the location information, candidate regions of the target vehicle are extracted from the visual data; The candidate regions are sequentially passed through the multiple convolutional blocks to extract features, thereby obtaining deep features; The depth features are classified using the fully connected layer to obtain the orientation and shooting distance of the target vehicle.

[0014] The advantages of adopting the above optional approach are: further utilizing convolutional blocks and fully connected layers to construct a vehicle orientation recognition model, automatically extracting and classifying candidate region depth features, achieving accurate identification of vehicle orientation and shooting distance, and providing a reliable orientation parameter basis for photo guidance.

[0015] In one alternative approach, the step of identifying the visual data using an information recognition algorithm to obtain the license plate and VIN code of the target vehicle includes: Based on the location information, the license plate area and VIN code area are located from the visual data; Using the information recognition algorithm, character recognition processing is performed on the license plate area and the VIN code area respectively to obtain the license plate and the VIN code; The license plate and the VIN code are subjected to format verification processing to verify their compliance.

[0016] The advantages of adopting the above-mentioned optional methods are: further locating the license plate and VIN code area through location information and performing character recognition and format verification, automatically verifying the compliance of vehicle identity information, avoiding manual input errors, ensuring data accuracy, and improving the reliability of the claims process.

[0017] In one alternative approach, the step of determining the accident type based on vehicle insurance claim information and obtaining the guidance rule corresponding to the accident type includes: Extract the number of vehicles involved in the accident from the vehicle insurance claim information; The accident type is determined based on the value of the quantity field; Based on the accident type, the corresponding guidance rule is retrieved from the predefined rule database.

[0018] The advantages of adopting the above optional methods are: further extracting the vehicle quantity field from the car insurance claim information to automatically determine the accident type, and calling the corresponding guidance rules to realize the automation and differentiated processing of rule matching, thereby enhancing the intelligent adaptability to different accident scenarios.

[0019] In one optional approach, the step of outputting photo-taking guidance information based on the location information, the orientation, the shooting distance, the license plate, the VIN code, and the guidance rules includes: When the accident type is a single-vehicle accident, the vehicle model information is determined based on the VIN code, and the location prompt, quantity verification and distance suggestion containing the vehicle model information are output. When the accident type is a multi-vehicle accident, the vehicle identification is determined based on the license plate, and the location prompt, the quantity verification and the distance suggestion containing the vehicle identification are output. Specifically, the location prompt is generated based on the location information and the photo requirements in the guidance rules, combined with the orientation and the shooting distance; the quantity verification is generated based on the list of required photos in the guidance rules, combined with the number of photos already taken; and the distance suggestion is generated based on the shooting specifications in the guidance rules.

[0020] The advantages of adopting the above-mentioned optional methods are: further distinguishing between single-vehicle accidents and multi-vehicle accidents, generating personalized guidance information based on VIN codes or license plates, including location prompts, quantity verification, and distance suggestions, making photo guidance more targeted and improving the standardization and efficiency of the investigation operation.

[0021] In one alternative approach, the step of updating the shooting progress in real time through the mobile terminal's visual interface during the output of the shooting guidance information, and distinguishing between incomplete and completed shooting locations through color changes in the visual interface, includes: A preset orientation legend is displayed in the visual interface of the mobile terminal; Based on the directional prompts, the incomplete shooting locations are marked in the directional legend with a first color, and based on the quantity verification, the completed shooting locations are marked in the directional legend with a second color. The labeling status of the first and second colors in the orientation legend is updated in real time.

[0022] The advantages of adopting the above-mentioned optional methods are: further displaying the incomplete and completed shooting status in real time through color-changing directional legends on the mobile terminal visualization interface, providing intuitive progress feedback, helping surveyors to quickly grasp the task completion status, avoiding omissions and optimizing user experience.

[0023] Secondly, this invention provides a vehicle insurance survey photo guidance system based on multi-model fusion, the technical solution of which is as follows: The vehicle detection module is used to detect visual data containing the target vehicle collected by the mobile terminal based on the vehicle detection model, and obtain the location information of the target vehicle. The orientation recognition module is used to identify the location information using a vehicle orientation recognition model to obtain the orientation and shooting distance of the target vehicle; The information recognition module is used to recognize the visual data through an information recognition algorithm to obtain the license plate and VIN code of the target vehicle; The rule acquisition module is used to determine the accident type based on the vehicle insurance claim information and acquire the guidance rules corresponding to the accident type. The photo guidance module is used to output photo guidance information based on the location information, the orientation, the shooting distance, the license plate, the VIN code, and the guidance rules; wherein, the photo guidance information includes: orientation prompts, quantity verification, and distance suggestions; The dynamic display module is used to update the shooting progress in real time through the visual interface of the mobile terminal during the output of the shooting guidance information, and to distinguish between the incomplete and completed shooting positions through color changes in the visual interface. The data association module is used to associate and store the image taken according to the photo-taking guidance information with the license plate and / or the VIN code.

[0024] The beneficial effects of the vehicle insurance survey photo guidance system based on multi-model fusion of the present invention are as follows: The system of this invention can solve the problems of difficult target positioning, non-standard operation and weak data correlation caused by the lack of photo guidance in traditional car insurance surveys, improve photo efficiency and photo quality, reduce labor costs and shorten the car insurance claims cycle.

[0025] Thirdly, the technical solution of an electronic device according to the present invention is as follows: It includes a memory, a processor, and a program stored in the memory and running on the processor, wherein the processor executes the program to implement the steps of the multi-model fusion-based vehicle insurance survey and photography guidance method of the present invention.

[0026] Fourthly, the technical solution of a computer-readable storage medium provided by the present invention is as follows: The computer-readable storage medium stores instructions that, when read, cause the computer-readable storage medium to perform the steps of the multi-model fusion-based vehicle insurance survey and photography guidance method of the present invention.

[0027] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0028] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a flowchart illustrating an embodiment of a vehicle insurance survey photo guidance method based on multi-model fusion according to the present invention. Figure 2 This is a schematic diagram illustrating the principle of multi-model fusion. Figure 3 This is a schematic diagram of an embodiment of a vehicle insurance survey and photo guidance system based on multi-model fusion according to the present invention; Figure 4 This is a schematic diagram of an embodiment of an electronic device according to the present invention. Detailed Implementation

[0029] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0030] Figure 1This diagram illustrates a flowchart of an embodiment of a multi-model fusion-based vehicle insurance claim inspection and photo guidance method provided by the present invention. This method can be executed by electronic devices such as terminal devices or servers. The terminal device can be any fixed or mobile terminal, such as user equipment (UE), mobile device, user terminal, terminal, cellular phone, cordless phone, personal digital assistant (PDA), handheld device, computing device, in-vehicle device, or wearable device. The server can be a single server or a server cluster consisting of multiple servers. Any electronic device can implement the multi-model fusion-based vehicle insurance claim inspection and photo guidance method by having its processor call computer-readable instructions stored in its memory. Figure 1 As shown, it includes the following steps: S1. Based on the vehicle detection model, the visual data containing the target vehicle collected by the mobile terminal is detected to obtain the location information of the target vehicle.

[0031] The vehicle detection model refers to a lightweight neural network model based on computer vision, used to identify and locate vehicles from visual data. For example, an investigator uses a smartphone to photograph an accident scene, and the vehicle detection model analyzes the image and outputs the bounding box of the car's location. The mobile terminal refers to a portable electronic device with computing, storage, and display capabilities. For example, an investigator's smartphone is used to capture vehicle images and run a photo-guided application. The target vehicle refers to the vehicle involved in the accident that needs to be photographed during the car insurance investigation. For example, a car damaged in an accident is the subject of the investigation. Visual data refers to digital images or video streams containing vehicle information collected by image sensors. For example, a scene photo containing the car taken by a smartphone camera. Location information refers to the spatial coordinate range of the vehicle in the visual data. For example, the vehicle detection model outputs the bounding box coordinates (x_min, y_min, x_max, y_max) of the car in the image.

[0032] S2. Use a vehicle orientation recognition model to identify the location information to obtain the orientation and shooting distance of the target vehicle.

[0033] The vehicle orientation recognition model refers to a deep learning model used to determine the relative orientation and shooting distance of a vehicle. For example, after analyzing an image of a car, the model identifies its orientation as 45° to the left front and its shooting distance as close-up. Orientation refers to the vehicle's position relative to the shooting angle; for example, the model outputs that the car's orientation is directly behind. Shooting distance refers to the relative distance between the camera and the vehicle; for example, if the model determines the shooting distance to be medium shot, it means the camera is approximately 5-10 meters away from the car.

[0034] S3. The visual data is identified using an information recognition algorithm to obtain the license plate and VIN code of the target vehicle.

[0035] Information recognition algorithms refer to computer algorithms used to extract specific character information from visual data. For example, an information recognition algorithm might identify the license plate number "ABC123" and the VIN code "1HGCM82633A123456" from a car image. A license plate is a vehicle's identification plate; for example, the information recognition algorithm extracts the license plate characters "ABC123" from a car image. A VIN code is a vehicle identification number, a unique identifier for a vehicle; for example, the information recognition algorithm identifies the VIN code "1HGCM82633A123456" from an image of a car's windshield.

[0036] S4. Determine the accident type based on the vehicle insurance claim information and obtain the guidance rules corresponding to the accident type.

[0037] Among these, "auto insurance claim information" refers to textual or structured information recording basic accident data in insurance claim cases; for example, the claim information may include information indicating that the accident involved one vehicle. "Accident type" refers to the accident category categorized based on the number of vehicles involved; for example, the accident type may be determined to be a single-vehicle accident based on the auto insurance claim information. "Guidance rules" refers to predefined photographing specifications and requirements; for example, calling guidance rules for single-vehicle accidents from a predefined rule database may require taking photos of the four corners of the vehicle and close-up photos of the damage.

[0038] S5. Based on the location information, orientation, shooting distance, license plate, VIN code, and guidance rules, output photo-taking guidance information; wherein, the photo-taking guidance information includes: orientation prompts, quantity verification, and distance suggestions.

[0039] The photo guidance information refers to: real-time generated shooting instructions; for example, the photo guidance app outputs guidance information including directional prompts "45° left front," quantity verification "1 more left front shot needed," and distance suggestion "close-up." Directional prompts refer to: textual or graphic instructions on the shooting angle; for example, the visual interface displays "Please shoot at a 45° left front." Quantity verification refers to: prompts to check the number of photos already taken and those yet to be taken; for example, after verification, the photo guidance app prompts "3 corner shots completed, 1 more right rear shot needed." Distance suggestion refers to: recommended shooting distances; for example, the photo guidance app suggests "Please shoot close-up photos" based on guidance rules.

[0040] S6. During the process of outputting the photo-taking guidance information, the photo-taking progress is updated in real time through the visual interface of the mobile terminal, and the incomplete and completed shooting positions are distinguished by color changes in the visual interface.

[0041] The visual interface refers to a user interface on a mobile terminal that displays graphical elements; for example, a camera interface on a smartphone screen that displays a location map and color progress indicators. Camera progress refers to the dynamic status of the shooting task; for example, a visual interface might use color changes to indicate that 5 out of 8 directions have been completed. Color changes refer to the dynamic updating of the colors of interface elements; for example, incomplete directions are displayed in yellow, and completed directions turn green. Shooting direction refers to the standardized vehicle shooting angle; for example, the 8 directions defined in guidance rules include directly in front, 45° to the left front, and directly to the left.

[0042] S7. The image taken according to the photo-taking guidance information is associated with and stored with the license plate and / or the VIN code.

[0043] Specifically, after taking images based on the photo-taking guidance information, the captured image file is mapped to the license plate, VIN code, or both. Through database operations, the image file path and the license plate, VIN code, or a combination of both are stored as a record in the case database, thereby realizing the automatic association and persistent storage of images and vehicle identity information.

[0044] The technical solution of this embodiment can solve the problems of difficult target positioning, non-standard operation and weak data correlation caused by the lack of photo guidance in traditional auto insurance surveys, improve photo efficiency and photo quality, reduce labor costs and shorten the auto insurance claims cycle.

[0045] In one alternative approach, the vehicle detection model includes: a feature extraction network, a feature pyramid network, and a detection head; S1 specifically includes: The mobile terminal acquires the visual data containing the target vehicle in real time.

[0046] The visual data is input into the feature extraction network of the vehicle detection model to extract multi-scale feature maps of the visual data.

[0047] In this context, a feature extraction network refers to a neural network structure in a vehicle detection model used to extract visual features layer by layer from an input image. For example, after receiving an image of a car, the feature extraction network extracts features step by step through multiple convolutional and pooling layers. The first convolutional layer identifies edge and color features, the middle convolutional layer identifies features of components such as wheels and windows, and the higher-level convolutional layer identifies the overall vehicle outline features, ultimately outputting a multi-scale feature map containing different levels of abstraction. A multi-scale feature map refers to a feature representation extracted from different levels of visual data; for example, the feature extraction network outputs multiple feature maps containing detailed and semantic information.

[0048] The multi-scale feature map is input into the feature pyramid network for feature fusion to generate enhanced multi-scale features.

[0049] In this context, the Feature Pyramid Network refers to a feature enhancement structure used to fuse multi-scale features. For example, after receiving feature maps of multiple scales from the feature extraction network, the Feature Pyramid Network fuses high-level semantic features with low-level detailed features through a top-down path, and then aligns the feature maps of different levels through lateral connections, ultimately generating enhanced multi-scale features containing rich semantic information and precise localization information. Enhanced multi-scale features refer to features that have been fused and optimized by the Feature Pyramid Network; for example, the fused features are used to more accurately predict car bounding boxes.

[0050] The enhanced multi-scale features are processed by the detection head to regress and generate the bounding box coordinates of the target vehicle.

[0051] In this context, the detection head refers to the network component in the vehicle detection model that outputs the detection results. For example, after receiving enhanced multi-scale features from the feature pyramid network, the detection head processes the features through two parallel branches: the classification branch predicts whether a vehicle exists at each location and its confidence level, while the regression branch predicts the bounding box coordinates of the vehicle, ultimately outputting the precise location of the car in the image. The bounding box coordinates refer to the rectangular coordinates that define the vehicle's position in the image; for example, the detection head outputs the coordinates of the top-left corner (100, 150) and bottom-right corner (300, 400) of the car's bounding box.

[0052] The position information of the target vehicle in the visual data is determined based on the bounding box coordinates.

[0053] Among the above-mentioned optional approaches, a multi-scale feature fusion architecture consisting of a feature extraction network, a feature pyramid network, and a detection head is further used to enhance the vehicle detection model's adaptability to complex scenes and different shooting conditions, significantly improving the detection accuracy and robustness of target vehicle location information.

[0054] In one optional embodiment, the vehicle orientation recognition model includes: a plurality of convolutional blocks connected in sequence and a fully connected layer; S2 specifically includes: Based on the location information, candidate regions for the target vehicle are extracted from the visual data.

[0055] Candidate regions refer to the regions of interest for vehicles extracted from visual data; for example, regions containing the entire vehicle are extracted from car images based on location information.

[0056] The candidate regions are sequentially processed through the multiple convolutional blocks to extract features, thereby obtaining deep features.

[0057] In this context, multiple convolutional blocks refer to a group of convolutional layers connected sequentially in a vehicle location recognition model. For example, three convolutional blocks process vehicle candidate regions in sequence. The first convolutional block contains convolutional layers, batch normalization layers, and a ReLU activation function to extract basic features. The second convolutional block increases the number of channels to extract more complex features. The third convolutional block further refines the features, ultimately outputting a discriminative deep feature vector. Deep features refer to high-level feature representations extracted through multi-layer neural networks. For example, after processing candidate regions, convolutional blocks generate deep feature vectors for classification.

[0058] The depth features are classified using the fully connected layer to obtain the orientation and shooting distance of the target vehicle.

[0059] Among them, the fully connected layer refers to a neural network layer used to map feature vectors to classification results. For example, after receiving the depth feature vectors output by the convolutional block, the fully connected layer transforms the feature dimension into the number of categories through the weight matrix, and then outputs the probability values ​​of each orientation category and shooting distance category through the softmax activation function, finally determining that the vehicle's orientation is directly behind and the shooting distance is close-up.

[0060] In the above-mentioned optional methods, a vehicle orientation recognition model is further constructed by using convolutional blocks and fully connected layers to automatically extract and classify the depth features of candidate regions, thereby achieving accurate recognition of vehicle orientation and shooting distance, and providing a reliable orientation parameter basis for photo guidance.

[0061] In one alternative approach, S3 specifically includes: Based on the location information, the license plate area and VIN code area are located from the visual data.

[0062] The license plate area refers to the image area in the visual data that contains license plate characters; for example, a rectangular area located from a car image that contains the license plate "ABC123". The VIN code area refers to the image area in the visual data that contains VIN code characters; for example, the area near the windshield located from a car image that contains the VIN code "1HGCM82633A123456".

[0063] Using the aforementioned information recognition algorithm, character recognition processing is performed on the license plate area and the VIN code area respectively to obtain the license plate and the VIN code.

[0064] Character recognition refers to the process of extracting text information from an image region; for example, an information recognition algorithm performs character recognition on a license plate region and outputs the string "ABC123".

[0065] The license plate and the VIN code are subjected to format verification processing to verify their compliance.

[0066] Format verification refers to the process of verifying the recognition results according to rules; for example, the information recognition algorithm verifies whether the license plate "ABC123" conforms to the license plate coding standard.

[0067] Among the above-mentioned optional methods, the license plate and VIN code areas can be further located by using location information and character recognition and format verification can be performed to automatically verify the compliance of vehicle identity information, avoid manual input errors, ensure data accuracy and improve the reliability of the claims process.

[0068] It should be noted that, as Figure 2 As shown, the mobile terminal inputs visual data video stream through the application and processes it frame by frame by calling the core algorithm of the software development kit. When the tracker list is not empty, the KCF algorithm is used to predict the position information of the target vehicle in the current frame. When the tracker list is empty, the vehicle detection model (PicoDet) is called to detect vehicles to reduce processing frequency and save power. Then, the state of the successfully matched trackers is updated, trackers are created for new detection boxes, and trackers that have not been matched for a long time are discarded. For each target vehicle, candidate regions are extracted from the visual data based on the position information, and the vehicle orientation recognition model (PPLCNet) is called to identify the vehicle's orientation and shooting distance. Finally, the structured data containing position information, orientation, and shooting distance is returned to the application.

[0069] In one alternative approach, S4 specifically includes: Extract the number of vehicles involved in the accident from the car insurance claim information.

[0070] The term "vehicles involved in the accident" refers to the number of vehicles related to the accident recorded in the accident report. For example, a value of 1 in the "quantity" field of a car insurance claim indicates that one sedan is involved. The "quantity" field refers to the data item representing the number of vehicles in the accident report; for example, the "vehicle_count" field, extracted from structured accident reports, has a value of 1.

[0071] The accident type is determined based on the value of the quantity field.

[0072] The value of the quantity field refers to the specific number of vehicles involved in the accident recorded in the car insurance claim information; for example, if the value of the quantity field extracted from the car insurance claim information is 2, it means that the accident involved two cars.

[0073] Based on the accident type, the corresponding guidance rule is retrieved from the predefined rule database.

[0074] The predefined rule database refers to a data set that stores guidance rules for different accident types; for example, the predefined rule database contains rules that require taking four-corner photos for single-vehicle accidents and photos from all directions for multi-vehicle accidents.

[0075] Among the above optional methods, the number of vehicles can be extracted from the car insurance claim information to automatically determine the accident type and call the corresponding guidance rules to realize the automation and differentiated processing of rule matching, thereby enhancing the intelligent adaptability to different accident scenarios.

[0076] In one alternative approach, S5 specifically includes: When the accident type is a single-vehicle accident, the vehicle model information is determined based on the VIN code, and the location prompt, quantity verification, and distance suggestion containing the vehicle model information are output. Vehicle model information refers to the vehicle model data obtained through VIN code parsing; for example, the VIN code "1HGCM82633A123456" retrieves the sedan model as "a certain brand and model".

[0077] Among them, a single-vehicle accident refers to an accident involving only one vehicle; for example, an accident in which a car crashes into a guardrail on its own is judged as a single-vehicle accident.

[0078] When the accident type is a multi-vehicle accident, the vehicle identification is determined based on the license plate, and the location prompt, quantity verification and distance suggestion containing the vehicle identification are output.

[0079] Multi-vehicle accidents refer to accidents involving two or more vehicles; for example, a quantity field value of 2 in the accident report indicates a two-vehicle collision. Vehicle identification refers to information used to distinguish different vehicles; for example, in a multi-vehicle accident, a specific car is identified by its license plate "ABC123".

[0080] Specifically, the location prompt is generated based on the location information and the photo requirements in the guidance rules, combined with the orientation and the shooting distance; the quantity verification is generated based on the list of required photos in the guidance rules, combined with the number of photos already taken; and the distance suggestion is generated based on the shooting specifications in the guidance rules.

[0081] The photography requirements refer to the specific regulations regarding orientation and distance in the guidance rules; for example, the guidance rules for single-vehicle accidents require shooting from the left front 45° and right front 45° angles. The list of required photos refers to the list of mandatory shooting positions and quantities; for example, the guidance rules stipulate one photo from each of the four corners and at least three close-up photos of the damage. Shooting specifications refer to the technical requirements for shooting distance and angle; for example, the guidance rules stipulate that close-up photos must clearly show the details of the damage.

[0082] Among the above-mentioned optional methods, we can further distinguish between single-vehicle accidents and multi-vehicle accidents, and generate personalized guidance information based on VIN codes or license plates, including location prompts, quantity verification and distance suggestions, so as to make the photo guidance more targeted and improve the standardization and efficiency of the investigation operation.

[0083] In one alternative approach, S6 specifically includes: A preset orientation map is displayed in the visual interface of the mobile terminal.

[0084] The preset orientation legend refers to the pre-designed orientation diagram in the visualization interface; for example, the bottom of the visualization interface displays a vehicle outline diagram containing 8 standard orientations.

[0085] Based on the directional prompts, incomplete shooting locations are marked in the directional legend using a first color, and based on the quantity verification, completed shooting locations are marked in the directional legend using a second color.

[0086] The first color refers to the color used to indicate an incomplete state; for example, in a directional map legend, an incomplete direction is displayed as yellow. The second color refers to the color used to indicate a completed state; for example, in a directional map legend, a completed direction is displayed as green.

[0087] The labeling status of the first and second colors in the orientation legend is updated in real time.

[0088] Among the above-mentioned optional methods, the incomplete and completed shooting status can be displayed in real time on the mobile terminal visual interface through a color-changing directional map, providing intuitive progress feedback, helping surveyors quickly grasp the task completion status, avoiding omissions and optimizing the user experience.

[0089] To better illustrate the technical solution of this embodiment, the following example is used for complete explanation: S10. Collect visual data containing the target vehicle through the mobile terminal camera; S20. Based on the vehicle detection model, the visual data is processed to obtain the location information of the target vehicle. The vehicle detection model includes a feature extraction network, a feature pyramid network, and a detection head. The feature extraction network extracts multi-scale feature maps from the visual data. The feature pyramid network performs feature fusion on the multi-scale feature maps to generate enhanced multi-scale features. The detection head processes the enhanced multi-scale features and regresses to generate the bounding box coordinates of the target vehicle. The location information of the target vehicle in the visual data is determined based on the bounding box coordinates. S30. Use a vehicle orientation recognition model to identify the location information and obtain the orientation and shooting distance of the target vehicle. The vehicle orientation recognition model includes multiple convolutional blocks connected in sequence and a fully connected layer. Based on the location information, the candidate region of the target vehicle is extracted from the visual data. The candidate region is then processed through multiple convolutional blocks to extract the depth features. The depth features are then classified through the fully connected layer to obtain the orientation and shooting distance of the target vehicle. S40. The visual data is identified by the information recognition algorithm to obtain the license plate and VIN code of the target vehicle. The license plate area and VIN code area are located from the visual data based on the location information. The information recognition algorithm is used to perform character recognition processing on the license plate area and VIN code area respectively to obtain the license plate and VIN code. The format verification processing is performed on the license plate and VIN code respectively to verify the compliance of the license plate and VIN code. S50. Determine the accident type based on the vehicle insurance claim information and obtain the corresponding guidance rules for the accident type. Extract the number of vehicles involved in the accident from the vehicle insurance claim information. Determine the accident type as a single-vehicle accident based on the value of the number field. Call the corresponding guidance rules from the predefined rule database based on the accident type. S60. Based on location information, orientation, shooting distance, license plate, VIN code, and guidance rules, output photo-taking guidance information, which includes orientation prompts, quantity verification, and distance suggestions. When the accident type is a single-vehicle accident, the vehicle model information is determined based on the VIN code, and orientation prompts, quantity verification, and distance suggestions containing vehicle model information are output. Orientation prompts are generated based on the location information and the photo-taking requirements in the guidance rules, combined with orientation and shooting distance. Quantity verification is generated based on the list of required photos in the guidance rules, combined with the number of photos already taken. Distance suggestions are generated based on the shooting specifications in the guidance rules. S70. During the process of outputting photo-taking guidance information, the photo-taking progress is updated in real time through the visual interface of the mobile terminal. The incomplete and completed shooting locations are distinguished by color changes in the visual interface. A preset location legend is displayed in the visual interface of the mobile terminal. According to the location prompts, the incomplete shooting locations are marked in the location legend with the first color. According to the quantity verification, the completed shooting locations are marked in the location legend with the second color. The marking status of the first and second colors in the location legend is updated in real time. S80. The image taken according to the photo-taking guidance information is associated with the license plate and VIN code and stored. After the image is taken according to the photo-taking guidance information, the image file is mapped to the license plate and VIN code. The image file path and the license plate and VIN code are stored as a record in the case database through database operations.

[0090] Figure 3 This diagram illustrates the structure of an embodiment of a multi-model fusion-based vehicle insurance survey and photo guidance system 200 provided by the present invention. Figure 3 As shown, the vehicle insurance survey photo guidance system 200 based on multi-model fusion includes: The vehicle detection module 201 is used to detect visual data containing the target vehicle collected by the mobile terminal based on the vehicle detection model, and obtain the location information of the target vehicle. The orientation recognition module 202 is used to identify the location information using a vehicle orientation recognition model to obtain the orientation and shooting distance of the target vehicle; The information recognition module 203 is used to recognize the visual data through an information recognition algorithm to obtain the license plate and VIN code of the target vehicle; The rule acquisition module 204 is used to determine the accident type based on the vehicle insurance claim information and acquire the guidance rule corresponding to the accident type. The photo guidance module 205 is used to output photo guidance information based on the location information, the orientation, the shooting distance, the license plate, the VIN code, and the guidance rules; wherein, the photo guidance information includes: orientation prompts, quantity verification, and distance suggestions; The dynamic display module 206 is used to update the shooting progress in real time through the visual interface of the mobile terminal during the process of outputting the shooting guidance information, and to distinguish between the incomplete and completed shooting positions in the visual interface by color change. The data association module 207 is used to associate and store the image taken according to the photo-taking guidance information with the license plate and / or the VIN code.

[0091] In one alternative embodiment, the vehicle detection model includes: a feature extraction network, a feature pyramid network, and a detection head; the vehicle detection module 201 is specifically used for: The mobile terminal acquires the visual data containing the target vehicle in real time. The visual data is input into the feature extraction network of the vehicle detection model to extract multi-scale feature maps of the visual data. The multi-scale feature map is input into the feature pyramid network for feature fusion to generate enhanced multi-scale features; The enhanced multi-scale features are processed by the detection head to regress and generate the bounding box coordinates of the target vehicle. The position information of the target vehicle in the visual data is determined based on the bounding box coordinates.

[0092] In one optional embodiment, the vehicle orientation recognition model includes: a plurality of convolutional blocks connected in sequence and a fully connected layer; the orientation recognition module 202 is specifically used for: Based on the location information, candidate regions of the target vehicle are extracted from the visual data; The candidate regions are sequentially passed through the multiple convolutional blocks to extract features, thereby obtaining deep features; The depth features are classified using the fully connected layer to obtain the orientation and shooting distance of the target vehicle.

[0093] In one alternative embodiment, the information recognition module 203 is specifically used for: Based on the location information, the license plate area and VIN code area are located from the visual data; Using the information recognition algorithm, character recognition processing is performed on the license plate area and the VIN code area respectively to obtain the license plate and the VIN code; The license plate and the VIN code are subjected to format verification processing to verify their compliance.

[0094] In an alternative embodiment, the rule acquisition module 204 is specifically used for: Extract the number of vehicles involved in the accident from the vehicle insurance claim information; The accident type is determined based on the value of the quantity field; Based on the accident type, the corresponding guidance rule is retrieved from the predefined rule database.

[0095] In one alternative embodiment, the photo-taking guidance module 205 is specifically used for: When the accident type is a single-vehicle accident, the vehicle model information is determined based on the VIN code, and the location prompt, quantity verification and distance suggestion containing the vehicle model information are output. When the accident type is a multi-vehicle accident, the vehicle identification is determined based on the license plate, and the location prompt, the quantity verification and the distance suggestion containing the vehicle identification are output. Specifically, the location prompt is generated based on the location information and the photo requirements in the guidance rules, combined with the orientation and the shooting distance; the quantity verification is generated based on the list of required photos in the guidance rules, combined with the number of photos already taken; and the distance suggestion is generated based on the shooting specifications in the guidance rules.

[0096] In an alternative embodiment, the dynamic display module 206 is specifically used for: A preset orientation legend is displayed in the visual interface of the mobile terminal; Based on the directional prompts, the incomplete shooting locations are marked in the directional legend with a first color, and based on the quantity verification, the completed shooting locations are marked in the directional legend with a second color. The labeling status of the first and second colors in the orientation legend is updated in real time.

[0097] It should be noted that the beneficial effects of the multi-model fusion-based vehicle insurance inspection and photography guidance system 200 provided in the above embodiments are the same as those of the multi-model fusion-based vehicle insurance inspection and photography guidance method described above, and will not be repeated here. Furthermore, the system provided in the above embodiments is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the system can be divided into different functional modules according to the actual situation to complete all or part of the functions described above. In addition, the system and method embodiments provided in the above embodiments belong to the same concept, and their specific implementation process is detailed in the method embodiments, and will not be repeated here.

[0098] The multi-model fusion-based vehicle insurance inspection and photography guidance system 200 of the present invention can be a computer program (including program code) running on a computer device. For example, the multi-model fusion-based vehicle insurance inspection and photography guidance system 200 of the present invention is an application software that can be used to execute the corresponding steps in the multi-model fusion-based vehicle insurance inspection and photography guidance method of the present invention.

[0099] In some embodiments, the multi-model fusion-based vehicle insurance inspection and photography guidance system 200 of the present invention can be implemented in a combination of hardware and software. As an example, the multi-model fusion-based vehicle insurance inspection and photography guidance system 200 of the present invention can be a processor in the form of a hardware decoding processor, which is programmed to execute the multi-model fusion-based vehicle insurance inspection and photography guidance method of the present invention. For example, the processor in the form of a hardware decoding processor can be one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic components.

[0100] The modules described in the embodiments of this invention can be implemented in software or hardware. The names of the modules are not, in some cases, limiting the scope of the module itself.

[0101] An electronic device according to an embodiment of the present invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the above-mentioned vehicle insurance inspection and photography guidance methods based on multi-model fusion. That is, an electronic device according to an embodiment of the present invention may include, but is not limited to: a processor and a memory; the memory is used to store the computer program; the processor is used to execute the vehicle insurance inspection and photography guidance method based on multi-model fusion shown in any embodiment of the present invention by calling the computer program.

[0102] In one alternative embodiment, an electronic device is provided, such as Figure 4 As shown, Figure 4 The illustrated electronic device 4000 includes a processor 4001 and a memory 4003. The processor 4001 and the memory 4003 are connected, for example, via a bus 4002. Optionally, the electronic device 4000 may further include a transceiver 4004, which can be used for data interaction between the electronic device and other electronic devices, such as sending and / or receiving data. It should be noted that in practical applications, the transceiver 4004 is not limited to one type, and the structure of the electronic device 4000 does not constitute a limitation on the embodiments of the present invention.

[0103] Processor 4001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this invention. Processor 4001 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0104] Bus 4002 may include a path for transmitting information between the aforementioned components. Bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Bus 4002 can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 4 The bus 4002 is represented by only one thick line, but this does not mean that there is only one bus or one type of bus.

[0105] The memory 4003 may be ROM (Read Only Memory) or other types of static storage devices capable of storing static information and instructions, RAM (Random Access Memory) or other types of dynamic storage devices capable of storing information and instructions, or EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto.

[0106] The memory 4003 stores application code (computer program) for executing the present invention, and its execution is controlled by the processor 4001. The processor 4001 executes the application code stored in the memory 4003 to implement the content shown in the foregoing method embodiments.

[0107] Among them, electronic devices can also be terminal devices. A terminal device can be any terminal device that can install applications and access web pages through applications, including at least one of smartphones, tablets, laptops, desktop computers, smart speakers, smartwatches, smart TVs, and smart in-vehicle devices.

[0108] It should be noted that, Figure 4 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0109] An embodiment of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements any of the above-mentioned vehicle insurance survey and photography guidance methods based on multi-model fusion.

[0110] Alternatively, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), magnetic tape, a floppy disk, and an optical data storage device, etc.

[0111] In an exemplary embodiment, a computer program product or computer program is also provided, which includes computer instructions stored in a computer-readable storage medium. The processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the electronic device to perform the aforementioned multi-model fusion-based vehicle insurance survey and photography guidance method.

[0112] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0113] It should be understood that the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of methods and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0114] The computer-readable storage medium provided in this invention can be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0115] The aforementioned computer-readable storage medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the method shown in the above embodiments.

[0116] The above description is merely a preferred embodiment of the present invention and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this invention is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this invention.

[0117] It should be noted that the terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and represent a limitation on a specific order or sequence. Where appropriate, the order of use for similar objects can be interchanged so that the embodiments of this application described herein can be implemented in an order other than that shown or described.

[0118] Those skilled in the art will recognize that this invention can be implemented as a system, method, or computer program product. Therefore, this invention can be specifically implemented in the following forms: it can be entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software, generally referred to herein as a "circuit," "module," or "system." Furthermore, in some embodiments, this invention can also be implemented as a computer program product contained in one or more computer-readable media, which includes computer-readable program code.

[0119] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for guiding vehicle insurance claims investigation through photography based on multi-model fusion, characterized in that, include: Based on the vehicle detection model, visual data containing the target vehicle collected through a mobile terminal is detected to obtain the location information of the target vehicle; The location information is identified using a vehicle orientation recognition model to obtain the orientation and shooting distance of the target vehicle; The visual data is identified using an information recognition algorithm to obtain the license plate and VIN code of the target vehicle. Determine the accident type based on the car insurance claim information and obtain the corresponding guidance rules for the accident type; Based on the location information, orientation, shooting distance, license plate, VIN code, and guidance rules, photo-taking guidance information is output; wherein, the photo-taking guidance information includes: orientation prompts, quantity verification, and distance suggestions; During the process of outputting the photo-taking guidance information, the photo-taking progress is updated in real time through the visual interface of the mobile terminal, and the incomplete and completed shooting locations are distinguished by color changes in the visual interface. The image taken according to the photo-taking guidance information will be associated with and stored with the license plate and / or the VIN code.

2. The vehicle insurance survey photo guidance method based on multi-model fusion according to claim 1, characterized in that, The vehicle detection model includes: a feature extraction network, a feature pyramid network, and a detection head; the step of detecting visual data containing the target vehicle collected by a mobile terminal based on the vehicle detection model to obtain the location information of the target vehicle includes: The mobile terminal acquires the visual data containing the target vehicle in real time. The visual data is input into the feature extraction network of the vehicle detection model to extract multi-scale feature maps of the visual data. The multi-scale feature map is input into the feature pyramid network for feature fusion to generate enhanced multi-scale features; The enhanced multi-scale features are processed by the detection head to regress and generate the bounding box coordinates of the target vehicle. The position information of the target vehicle in the visual data is determined based on the bounding box coordinates.

3. The vehicle insurance survey photo guidance method based on multi-model fusion according to claim 1, characterized in that, The vehicle orientation recognition model includes: multiple convolutional blocks connected in sequence and a fully connected layer; the step of using the vehicle orientation recognition model to identify the location information and obtain the orientation and shooting distance of the target vehicle includes: Based on the location information, candidate regions of the target vehicle are extracted from the visual data; The candidate regions are sequentially passed through the multiple convolutional blocks to extract features, thereby obtaining deep features; The depth features are classified using the fully connected layer to obtain the orientation and shooting distance of the target vehicle.

4. The vehicle insurance survey photo guidance method based on multi-model fusion according to claim 1, characterized in that, The step of identifying the visual data using an information recognition algorithm to obtain the license plate and VIN code of the target vehicle includes: Based on the location information, the license plate area and VIN code area are located from the visual data; Using the information recognition algorithm, character recognition processing is performed on the license plate area and the VIN code area respectively to obtain the license plate and the VIN code; The license plate and the VIN code are subjected to format verification processing to verify their compliance.

5. The vehicle insurance survey photo guidance method based on multi-model fusion according to any one of claims 1 to 4, characterized in that, The steps of determining the accident type based on the vehicle insurance claim information and obtaining the corresponding guidance rules for the accident type include: Extract the number of vehicles involved in the accident from the vehicle insurance claim information; The accident type is determined based on the value of the quantity field; Based on the accident type, the corresponding guidance rule is retrieved from the predefined rule database.

6. The vehicle insurance survey photo guidance method based on multi-model fusion according to claim 5, characterized in that, The step of outputting photo-taking guidance information based on the location information, orientation, shooting distance, license plate, VIN code, and guidance rules includes: When the accident type is a single-vehicle accident, the vehicle model information is determined based on the VIN code, and the location prompt, quantity verification and distance suggestion containing the vehicle model information are output. When the accident type is a multi-vehicle accident, the vehicle identification is determined based on the license plate, and the location prompt, the quantity verification and the distance suggestion containing the vehicle identification are output. Specifically, the location prompt is generated based on the location information and the photo requirements in the guidance rules, combined with the orientation and the shooting distance; the quantity verification is generated based on the list of required photos in the guidance rules, combined with the number of photos already taken; and the distance suggestion is generated based on the shooting specifications in the guidance rules.

7. The vehicle insurance survey photo guidance method based on multi-model fusion according to claim 6, characterized in that, The step of updating the shooting progress in real time through the visual interface of the mobile terminal during the output of the shooting guidance information, and distinguishing between incomplete and completed shooting positions through color changes in the visual interface, includes: A preset orientation legend is displayed in the visual interface of the mobile terminal; Based on the directional prompts, the incomplete shooting locations are marked in the directional legend with a first color, and based on the quantity verification, the completed shooting locations are marked in the directional legend with a second color. The labeling status of the first and second colors in the orientation legend is updated in real time.

8. A vehicle insurance survey photo guidance system based on multi-model fusion, characterized in that, include: The vehicle detection module is used to detect visual data containing the target vehicle collected by the mobile terminal based on the vehicle detection model, and obtain the location information of the target vehicle. The orientation recognition module is used to identify the location information using a vehicle orientation recognition model to obtain the orientation and shooting distance of the target vehicle; The information recognition module is used to recognize the visual data through an information recognition algorithm to obtain the license plate and VIN code of the target vehicle; The rule acquisition module is used to determine the accident type based on the vehicle insurance claim information and acquire the guidance rules corresponding to the accident type. The photo guidance module is used to output photo guidance information based on the location information, the orientation, the shooting distance, the license plate, the VIN code, and the guidance rules; wherein, the photo guidance information includes: orientation prompts, quantity verification, and distance suggestions; The dynamic display module is used to update the shooting progress in real time through the visual interface of the mobile terminal during the output of the shooting guidance information, and to distinguish between the incomplete and completed shooting positions through color changes in the visual interface. The data association module is used to associate and store the image taken according to the photo-taking guidance information with the license plate and / or the VIN code.

9. An electronic device, characterized in that, The electronic device includes a processor coupled to a memory, the memory storing at least one computer program, which is loaded and executed by the processor to enable the electronic device to implement the vehicle insurance survey and photography guidance method based on multi-model fusion as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one computer program, which, when executed by a processor, implements the vehicle insurance survey and photography guidance method based on multi-model fusion as described in any one of claims 1 to 7.

Images