Vehicle insurance survey method and device
By using deep learning and multimodal large models to automatically process vehicle insurance survey information, the problem of low efficiency in vehicle insurance surveys has been solved, and rapid information integration and automated generation of survey reports have been achieved, thereby improving survey efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-25
- Publication Date
- 2026-04-03
AI Technical Summary
The efficiency of car insurance claims investigation is low, as the main sources of information rely on user self-entry and on-site collection by insurance company claims adjusters, resulting in low information collection efficiency.
By acquiring user information, accident scene photos, and document photos, deep learning and multimodal large models are used for image classification and information extraction to automatically generate investigation reports.
It enables rapid integration and automated processing of vehicle insurance survey information, improving survey efficiency and reducing waiting time for fragmented information collection.
Smart Images
Figure CN121788264A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of insurance technology, and in particular to a method and apparatus for vehicle insurance claims investigation. Background Technology
[0002] In the field of auto insurance, claims investigation is a crucial step in the claims process. Its core objective is to accurately collect accident-related information to provide a basis for subsequent damage assessment, verification, and compensation. Currently, the main sources of information for auto insurance claims investigation rely on user-entered data and on-site collection by insurance company investigators. This information then needs to be further reviewed by loss adjusters and verification personnel to complete the damage assessment and verification process, resulting in low efficiency in auto insurance claims investigation.
[0003] In conclusion, improving the efficiency of vehicle insurance claims investigation is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention
[0004] In view of this, this application provides a method and apparatus for vehicle insurance claims investigation, which aims to improve the efficiency of vehicle insurance claims investigation.
[0005] Firstly, this application provides a method for investigating vehicle insurance claims, including:
[0006] Obtain input information; the input information includes user information, photos of the accident scene, photos of the ID cards of the vehicle in question and the third-party vehicle, photos of the vehicle registration certificates of the vehicle in question and the third-party vehicle, photos of the driver's licenses of the vehicle in question and the third-party vehicle, and photos of the traffic accident liability certificate.
[0007] Classify all images in the input information to obtain image classification results;
[0008] Based on the image classification results, basic information is extracted; the basic information includes driver information and target vehicle information.
[0009] Based on the basic information and the image corresponding to the image classification result, accident information is extracted; the accident information includes the accident process, the authenticity analysis result, the cause of the accident, the location of the damage, and the estimated amount of damage.
[0010] Based on the aforementioned basic information and accident information, an investigation report is generated.
[0011] Optionally, classifying all images in the input information to obtain image classification results includes:
[0012] A deep learning algorithm is used to classify all images in the input information to obtain the image classification results; the image classification results include accident scene photos, vehicle identification number photos, ordinary cards and certificates, electronic cards and certificates, and traffic accident liability certificates.
[0013] Optionally, the step of extracting basic information based on the image classification result includes:
[0014] For the images corresponding to the accident scene photos category in the image classification results, a deep learning algorithm is used to extract the license plate number and license plate position, and a traditional image processing algorithm is used to extract the license plate color.
[0015] For the images corresponding to the vehicle identification number (VIN) photo category in the image classification results, a deep learning algorithm is used to extract the VIN.
[0016] For the images corresponding to the ordinary card / certificate category and the electronic card / certificate category in the image classification results, a deep learning algorithm is used to extract driver information and target vehicle information. The driver information includes name, document type, document number, driver's license number, and gender. The target vehicle information includes vehicle type, energy type, engine number, vehicle registration date, and vehicle issuance date.
[0017] Optionally, extracting accident information based on the basic information and the image corresponding to the image classification result includes:
[0018] The text information in the traffic accident report photo is extracted using the OCR model in the deep learning algorithm. The basic information and the user information are then input into the large language model to obtain the text information of the accident process.
[0019] The first scene photos of the accident and the basic information are input into a multimodal large model to obtain image information of the accident process;
[0020] The accident sequence is obtained by fusing textual information and image information about the accident sequence.
[0021] The basic information, the accident details, and the photos of the first accident scene are input into the multimodal large model to determine the rationality of the damage and obtain the authenticity analysis results.
[0022] The accident details and the photos of the first accident scene are input into the multimodal large model to obtain the cause of the accident;
[0023] The accident details and the photos of the first accident scene are input into the multimodal large model to obtain the location of the damage;
[0024] The accident details, the location of the damage, the photos of the accident scene, and business experience are input into a multimodal large model to obtain the estimated loss amount; the business experience is obtained by statistically analyzing the correspondence between parts, damage types, and compensation amounts in historical data.
[0025] Secondly, this application provides a vehicle insurance survey device, comprising:
[0026] The acquisition module is used to acquire input information, which includes user information, photos of the accident scene, photos of the ID cards of the vehicle in question and the third-party vehicle, photos of the vehicle registration certificates of the vehicle in question and the third-party vehicle, photos of the driver's licenses of the vehicle in question and the third-party vehicle, and photos of the traffic accident liability certificate.
[0027] The image classification module is used to classify all images in the input information and obtain image classification results;
[0028] The first extraction module is used to extract basic information based on the image classification results; the basic information includes driver information and target vehicle information.
[0029] The second extraction module is used to extract accident information based on the basic information and the image corresponding to the image classification result; the accident information includes the accident process, the authenticity analysis result, the cause of the accident, the location of the damage, and the estimated amount of damage.
[0030] The generation module is used to generate an investigation report based on the basic information and the accident information.
[0031] Optionally, the image classification module is specifically used for:
[0032] A deep learning algorithm is used to classify all images in the input information to obtain the image classification results; the image classification results include accident scene photos, vehicle identification number photos, ordinary cards and certificates, electronic cards and certificates, and traffic accident liability certificates.
[0033] Optionally, the first extraction module is specifically used for:
[0034] For the images corresponding to the accident scene photos category in the image classification results, a deep learning algorithm is used to extract the license plate number and license plate position, and a traditional image processing algorithm is used to extract the license plate color.
[0035] For the images corresponding to the vehicle identification number (VIN) photo category in the image classification results, a deep learning algorithm is used to extract the VIN.
[0036] For the images corresponding to the ordinary card / certificate category and the electronic card / certificate category in the image classification results, a deep learning algorithm is used to extract driver information and target vehicle information. The driver information includes name, document type, document number, driver's license number, and gender. The target vehicle information includes vehicle type, energy type, engine number, vehicle registration date, and vehicle issuance date.
[0037] Optionally, the second extraction module is specifically used for:
[0038] The text information in the traffic accident report photo is extracted using the OCR model in the deep learning algorithm. The basic information and the user information are then input into the large language model to obtain the text information of the accident process.
[0039] The first scene photos of the accident and the basic information are input into a multimodal large model to obtain image information of the accident process;
[0040] The accident sequence is obtained by fusing textual information and image information about the accident sequence.
[0041] The basic information, the accident details, and the photos of the first accident scene are input into the multimodal large model to determine the rationality of the damage and obtain the authenticity analysis results.
[0042] The accident details and the photos of the first accident scene are input into the multimodal large model to obtain the cause of the accident;
[0043] The accident details and the photos of the first accident scene are input into the multimodal large model to obtain the location of the damage;
[0044] The accident details, the location of the damage, the photos of the accident scene, and business experience are input into a multimodal large model to obtain the estimated loss amount; the business experience is obtained by statistically analyzing the correspondence between parts, damage types, and compensation amounts in historical data.
[0045] Thirdly, embodiments of this application provide a computer device, including: 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 the vehicle insurance survey method as described in any of the embodiments of the first aspect of this application.
[0046] Fourthly, embodiments of this application provide a computer-readable storage medium storing instructions that, when executed on a terminal device, cause the terminal device to perform the vehicle insurance survey method as described in any of the embodiments of the first aspect of this application.
[0047] This application provides a method for vehicle insurance claims investigation. When executing the method, input information is first acquired, including user information, photos of the accident scene, photos of the ID cards of the insured vehicle and the third-party vehicle, photos of the vehicle registration certificates of the insured vehicle and the third-party vehicle, photos of the driver's licenses of the insured vehicle and the third-party vehicle, and a photo of the traffic accident liability certificate. Then, all images in the input information are classified to obtain image classification results. Next, based on the image classification results, basic information is extracted, including driver information and insured vehicle information. Then, based on the basic information and the images corresponding to the image classification results, accident information is extracted, including the accident details, the results of the authenticity analysis, the cause of the accident, the location of the damage, and the estimated damage amount. Finally, based on the basic information and the accident information, an investigation report is generated. In this way, by integrating all key information such as user information, accident scene photos, photos of both parties' documents, and traffic accident liability certificates at once, the time-consuming back-and-forth waiting caused by fragmented information collection in traditional surveys is eliminated. At the same time, through intelligent classification, effective images are quickly filtered, and core information such as driver information, vehicle information, accident details, damage location, and estimated damage amount are automatically extracted based on algorithms, thereby improving the efficiency of car insurance surveys. Attached Figure Description
[0048] To more clearly illustrate the technical solutions in this embodiment or the prior art, the drawings used in the description of the embodiment or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0049] Figure 1 A flowchart illustrating a vehicle insurance survey method provided in this application embodiment;
[0050] Figure 2 This is a schematic diagram of the structure of a vehicle insurance surveying device provided in an embodiment of this application;
[0051] Figure 3 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation
[0052] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. This application provides a method and apparatus for vehicle insurance claims investigation, relating to the field of insurance technology. The above are merely examples and do not limit the application field of the method and apparatus provided in this application.
[0053] In the field of auto insurance, claims investigation is a crucial step in the claims process. Its core objective is to accurately collect accident-related information to provide a basis for subsequent damage assessment, verification, and compensation. Currently, the main sources of information for auto insurance claims investigation rely on user-entered data and on-site collection by insurance company investigators. This information then needs to be further reviewed by loss adjusters and verification personnel to complete the damage assessment and verification process, resulting in low efficiency in auto insurance claims investigation.
[0054] The inventor, through research, proposed the technical solution of this application. First, input information is acquired, including user information, photos of the accident scene, photos of the ID cards of the vehicle in question and the third-party vehicle, photos of the vehicle registration certificates of the vehicle in question and the third-party vehicle, photos of the driver's licenses of the vehicle in question and the third-party vehicle, and a photo of the traffic accident liability certificate. Then, all images in the input information are classified to obtain image classification results. Next, based on the image classification results, basic information is extracted, including driver information and vehicle information. Then, based on the basic information and the images corresponding to the image classification results, accident information is extracted, including the accident details, the results of the authenticity analysis, the cause of the accident, the location of the damage, and the estimated damage amount. Finally, based on the basic information and the accident information, an investigation report is generated. In this way, by integrating all key information such as user information, accident scene photos, photos of both parties' documents, and traffic accident liability certificates at once, the time-consuming back-and-forth waiting caused by fragmented information collection in traditional surveys is eliminated. At the same time, through intelligent classification, effective images are quickly filtered, and core information such as driver information, vehicle information, accident details, damage location, and estimated damage amount are automatically extracted based on algorithms, thereby improving the efficiency of car insurance surveys.
[0055] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are only a part of the embodiments of the present application, and not all of them. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present application. It should be noted that, for ease of description, only the parts related to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features in the embodiments of the present application can be combined with each other.
[0056] See Figure 1 , Figure 1 A flowchart of a vehicle insurance survey method provided in this application embodiment includes:
[0057] S101: Obtain input information.
[0058] First, it is necessary to obtain input information, which includes, but is not limited to, user information, photos of the accident scene, photos of the ID cards of the infringing vehicle and the third-party vehicle, photos of the vehicle registration certificates of the infringing vehicle and the third-party vehicle, photos of the driver's licenses of the infringing vehicle and the third-party vehicle, and photos of the traffic accident liability certificate.
[0059] S102: Classify all images in the input information to obtain image classification results.
[0060] Next, all images in the input information are classified to obtain the image classification results, as follows:
[0061] A deep learning algorithm is used to classify all images in the input information to obtain image classification results. The image classification results include categories such as accident scene photos, vehicle identification number photos, ordinary cards and certificates, electronic cards and certificates, and traffic accident liability certificates.
[0062] S103: Extract basic information based on image classification results.
[0063] First, for the images corresponding to the accident scene photos category in the image classification results, a deep learning algorithm is used to extract the license plate number and license plate position, and a traditional image processing algorithm is used to extract the license plate color.
[0064] Next, for the images corresponding to the vehicle identification number (VIN) photo categories in the image classification results, a deep learning algorithm is used to extract the VIN.
[0065] Finally, for the images corresponding to the ordinary card and electronic card categories in the image classification results, a deep learning algorithm is used to extract driver information and target vehicle information. The driver information includes name, document type, document number, driver's license number, and gender. The target vehicle information includes vehicle type, energy type, engine number, vehicle registration date, and vehicle issuance date.
[0066] S104: Extract accident information based on the basic information and the image corresponding to the image classification results.
[0067] Accident information includes the details of the accident, the results of the accuracy analysis, the cause of the accident, the location of the damage, and the estimated amount of damage.
[0068] The method for extracting the accident details is as follows: The text information in the traffic accident report photo is extracted using the OCR model in the deep learning algorithm. The basic information and user information are then input into the large language model to obtain the text information of the accident details. The accident scene photo and basic information are then input into the multimodal large model to obtain the image information of the accident details. Finally, the text information and image information of the accident details are fused to obtain the accident details.
[0069] The method for determining the authenticity analysis results is as follows:
[0070] The information of the target vehicle, the details of the accident, and photos of the accident scene are input into a multimodal large model for authenticity analysis to determine the reasonableness of the damage.
[0071] The method for determining the cause of the accident is as follows:
[0072] The accident details and photos of the accident scene are input into a multimodal large model for accident cause analysis.
[0073] The location of the damage is determined as follows:
[0074] The accident details and photos of the accident scene are input into the multimodal large model to locate the damage.
[0075] The method for determining the estimated loss amount is as follows:
[0076] The accident details, damage location, photos of the accident scene, and business experience are input into a multimodal big data model to obtain the estimated loss amount. The business experience is obtained by statistically analyzing the correspondence between parts, damage types, and compensation amounts in historical data.
[0077] S105: Generate an investigation report based on basic information and accident information.
[0078] The survey report includes the following information:
[0079] Information on the driver of the vehicle in question (name, document type & number, driver's license number, gender);
[0080] Vehicle information (license plate number, license plate color, vehicle type, power type, engine number, VIN, cause of the accident, registration date and issuance date of the vehicle registration certificate)
[0081] Accident information (cause of the accident, cause of the accident, liability for the accident, liability ratio, vehicle driving status, accident details, estimated damage amount, and authenticity analysis).
[0082] Furthermore, the embodiments of this application also include model training steps: the deep learning model is fine-tuned using business data; in the large language model, only the damage location extraction module is fine-tuned using business data under SFT supervision, while the other large language models use open-source base models; all large language models can be selected for SFT supervision fine-tuning or RL reinforcement learning according to the actual situation; the traditional image processing algorithm includes the following steps: cropping the license plate image according to the license plate coordinates, converting the cropped license plate image to HSV space for denoising, and using the KMeans algorithm to extract the main color of the license plate.
[0083] In the embodiments provided in this application, input information is first acquired, including user information, photos of the accident scene, photos of the ID cards of the insured vehicle and the third-party vehicle, photos of the vehicle registration certificates of the insured vehicle and the third-party vehicle, photos of the driver's licenses of the insured vehicle and the third-party vehicle, and a photo of the traffic accident liability certificate. Then, all images in the input information are classified to obtain image classification results. Next, based on the image classification results, basic information is extracted, including driver information and insured vehicle information. Then, based on the basic information and the images corresponding to the image classification results, accident information is extracted, including the accident details, authenticity analysis results, accident cause, damage location, and estimated damage amount. Finally, based on the basic information and the accident information, an investigation report is generated. In this way, by integrating all key information such as user information, accident scene photos, photos of both parties' identification documents, and the traffic accident liability certificate at once, the time-consuming back-and-forth waiting caused by fragmented information collection in traditional investigations is eliminated. Simultaneously, intelligent classification quickly filters effective images, and algorithms automatically extract core content such as driver information, insured vehicle information, accident details, damage location, and estimated damage amount, thereby improving the efficiency of vehicle insurance investigations.
[0084] The above are some specific implementations of the vehicle insurance survey method provided in this application. Based on this, this application also provides a corresponding device. The device provided in this application will be described below from the perspective of functional modularity.
[0085] See Figure 2 , Figure 2 This is a schematic diagram of the structure of a vehicle insurance surveying device provided in an embodiment of this application. The vehicle insurance surveying device 200 includes:
[0086] The acquisition module 210 is used to acquire input information; the input information includes user information, photos of the accident scene, photos of the ID cards of the vehicle in question and the third-party vehicle, photos of the vehicle registration certificates of the vehicle in question and the third-party vehicle, photos of the driver's licenses of the vehicle in question and the third-party vehicle, and photos of the traffic accident liability certificate.
[0087] Image classification module 220 is used to classify all images in the input information to obtain image classification results;
[0088] The first extraction module 230 is used to extract basic information based on the image classification results; the basic information includes driver information and target vehicle information.
[0089] The second extraction module 240 is used to extract accident information based on the basic information and the image corresponding to the image classification result; the accident information includes the accident process, the authenticity analysis result, the cause of the accident, the location of the damage, and the estimated amount of damage.
[0090] The generation module 250 is used to generate an investigation report based on the basic information and the accident information.
[0091] Optionally, the image classification module 220 is specifically used for:
[0092] A deep learning algorithm is used to classify all images in the input information to obtain the image classification results; the image classification results include accident scene photos, vehicle identification number photos, ordinary cards and certificates, electronic cards and certificates, and traffic accident liability certificates.
[0093] Optionally, the first extraction module 230 is specifically used for:
[0094] For the images corresponding to the accident scene photos category in the image classification results, a deep learning algorithm is used to extract the license plate number and license plate position, and a traditional image processing algorithm is used to extract the license plate color.
[0095] For the images corresponding to the vehicle identification number (VIN) photo category in the image classification results, a deep learning algorithm is used to extract the VIN.
[0096] For the images corresponding to the ordinary card / certificate category and the electronic card / certificate category in the image classification results, a deep learning algorithm is used to extract driver information and target vehicle information. The driver information includes name, document type, document number, driver's license number, and gender. The target vehicle information includes vehicle type, energy type, engine number, vehicle registration date, and vehicle issuance date.
[0097] Optionally, the second extraction module 240 is specifically used for:
[0098] The text information in the traffic accident report photo is extracted using the OCR model in the deep learning algorithm. The basic information and the user information are then input into the large language model to obtain the text information of the accident process.
[0099] The first scene photos of the accident and the basic information are input into a multimodal large model to obtain image information of the accident process;
[0100] The accident sequence is obtained by fusing textual information and image information about the accident sequence.
[0101] The basic information, the accident details, and the photos of the first accident scene are input into the multimodal large model to determine the rationality of the damage and obtain the authenticity analysis results.
[0102] The accident details and the photos of the first accident scene are input into the multimodal large model to obtain the cause of the accident;
[0103] The accident details and the photos of the first accident scene are input into the multimodal large model to obtain the location of the damage;
[0104] The accident details, the location of the damage, the photos of the accident scene, and business experience are input into a multimodal large model to obtain the estimated loss amount; the business experience is obtained by statistically analyzing the correspondence between parts, damage types, and compensation amounts in historical data.
[0105] This application also provides corresponding devices and computer storage media for implementing the solutions provided in this application.
[0106] like Figure 3 As shown, computer device 01 is represented in the form of a general-purpose computing device. The components of computer device 01 may include, but are not limited to: one or more processors or processor units 03, system memory 08, and bus 04 connecting different system components (including system memory 08 and processor unit 03).
[0107] Bus 04 represents one or more of several bus architectures, including memory buses or memory controllers, peripheral buses, graphics acceleration ports, processors, or local buses using any of the various bus architectures. Examples of these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.
[0108] Computer device 01 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by computer device 01, including volatile and non-volatile media, removable and non-removable media.
[0109] System memory 08 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 09 and / or cache memory 10. Computer device 01 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 11 may be used to read and write non-removable, non-volatile magnetic media (…). Figure 3 Not shown; usually referred to as a "hard drive"). Although Figure 3As not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 04 via one or more data media interfaces. System memory 08 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of the present invention.
[0110] A program / utility 12 having a set (at least one) of program modules 13 may be stored, for example, in system memory 08. Such program modules 13 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 13 typically perform the functions and / or methods described in the embodiments of the present invention.
[0111] Computer device 01 can also communicate with one or more external devices 02 (e.g., keyboard, pointing device, display 07, etc.), and with one or more devices that enable a user to interact with the computer device 01, and / or with any device that enables the computer device 01 to communicate with one or more other computing devices (e.g., network card, modem, etc.). This communication can be performed through input / output (I / O) interface 06. Furthermore, computer device 01 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) through network adapter 05. Figure 3 As shown, network adapter 05 communicates with other modules of computer device 01 via bus 04. It should be understood that, although... Figure 3 As not shown in the diagram, it can be used in conjunction with computer device 01 with other hardware and / or software modules, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.
[0112] The processor unit 03 executes various functional applications and data processing by running programs stored in the system memory 08, such as implementing a vehicle insurance survey method provided in the embodiments of this application.
[0113] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0114] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that all or part of the steps in the methods of the above embodiments can be implemented by means of software plus a general-purpose hardware platform. Based on this understanding, the technical solution of this application can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as a read-only memory (ROM) / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, a server, or a network communication device such as a router) to execute the methods described in various embodiments or some parts of the embodiments of this application.
[0115] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on its differences from other embodiments. In particular, the apparatus embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.
[0116] The above description is merely an exemplary implementation of this application and is not intended to limit the scope of protection of this application.
Claims
1. A method for vehicle insurance claims investigation, characterized in that, include: Obtain input information; the input information includes user information, photos of the accident scene, photos of the ID cards of the vehicle in question and the third-party vehicle, photos of the vehicle registration certificates of the vehicle in question and the third-party vehicle, photos of the driver's licenses of the vehicle in question and the third-party vehicle, and photos of the traffic accident liability certificate. Classify all images in the input information to obtain image classification results; Based on the image classification results, extract basic information; The basic information includes driver information and target vehicle information; Based on the basic information and the image corresponding to the image classification result, accident information is extracted; The accident information includes the accident details, the results of the accuracy analysis, the cause of the accident, the location of the damage, and the estimated amount of damage. Based on the aforementioned basic information and accident information, an investigation report is generated.
2. The method according to claim 1, characterized in that, The process of classifying all images in the input information to obtain image classification results includes: A deep learning algorithm is used to classify all images in the input information to obtain the image classification results; the image classification results include accident scene photos, vehicle identification number photos, ordinary cards and certificates, electronic cards and certificates, and traffic accident liability certificates.
3. The method according to claim 2, characterized in that, The extraction of basic information based on the image classification results includes: For the images corresponding to the accident scene photos category in the image classification results, a deep learning algorithm is used to extract the license plate number and license plate position, and a traditional image processing algorithm is used to extract the license plate color. For the images corresponding to the vehicle identification number (VIN) photo category in the image classification results, a deep learning algorithm is used to extract the VIN. For the images corresponding to the ordinary card / certificate category and the electronic card / certificate category in the image classification results, a deep learning algorithm is used to extract driver information and target vehicle information. The driver information includes name, document type, document number, driver's license number, and gender. The target vehicle information includes vehicle type, energy type, engine number, vehicle registration date, and vehicle issuance date.
4. The method according to claim 1, characterized in that, The step of extracting accident information based on the basic information and the image corresponding to the image classification result includes: The text information in the traffic accident report photo is extracted using the OCR model in the deep learning algorithm. The basic information and the user information are then input into the large language model to obtain the text information of the accident process. The accident scene photos and the basic information are input into a multimodal large model to obtain image information of the accident process; The accident sequence is obtained by fusing textual information and image information about the accident sequence. The basic information, the accident details, and the photos of the first accident scene are input into the multimodal large model to determine the rationality of the damage and obtain the authenticity analysis results. The accident details and the photos of the first accident scene are input into the multimodal large model to obtain the cause of the accident; The accident details and the photos of the first accident scene are input into the multimodal large model to obtain the location of the damage; The accident details, the location of the damage, the photos of the accident scene, and business experience are input into a multimodal large model to obtain the estimated loss amount; the business experience is obtained by statistically analyzing the correspondence between parts, damage types, and compensation amounts in historical data.
5. A vehicle insurance survey device, characterized in that, include: The acquisition module is used to acquire input information, which includes user information, photos of the accident scene, photos of the ID cards of the vehicle in question and the third-party vehicle, photos of the vehicle registration certificates of the vehicle in question and the third-party vehicle, photos of the driver's licenses of the vehicle in question and the third-party vehicle, and photos of the traffic accident liability certificate. The image classification module is used to classify all images in the input information and obtain image classification results; The first extraction module is used to extract basic information based on the image classification results; The basic information includes driver information and target vehicle information; The second extraction module is used to extract accident information based on the basic information and the image corresponding to the image classification result; The accident information includes the accident details, the results of the accuracy analysis, the cause of the accident, the location of the damage, and the estimated amount of damage. The generation module is used to generate an investigation report based on the basic information and the accident information.
6. The apparatus according to claim 5, characterized in that, The image classification module is specifically used for: A deep learning algorithm is used to classify all images in the input information to obtain the image classification results; the image classification results include accident scene photos, vehicle identification number photos, ordinary cards and certificates, electronic cards and certificates, and traffic accident liability certificates.
7. The apparatus according to claim 6, characterized in that, The first extraction module is specifically used for: For the images corresponding to the accident scene photos category in the image classification results, a deep learning algorithm is used to extract the license plate number and license plate position, and a traditional image processing algorithm is used to extract the license plate color. For the images corresponding to the vehicle identification number (VIN) photo category in the image classification results, a deep learning algorithm is used to extract the VIN. For the images corresponding to the ordinary card / certificate category and the electronic card / certificate category in the image classification results, a deep learning algorithm is used to extract driver information and target vehicle information. The driver information includes name, document type, document number, driver's license number, and gender. The target vehicle information includes vehicle type, energy type, engine number, vehicle registration date, and vehicle issuance date.
8. The apparatus according to claim 5, characterized in that, The second extraction module is specifically used for: The text information in the traffic accident report photo is extracted using the OCR model in the deep learning algorithm. The basic information and the user information are then input into the large language model to obtain the text information of the accident process. The accident scene photos and the basic information are input into a multimodal large model to obtain image information of the accident process; The accident sequence is obtained by fusing textual information and image information about the accident sequence. The basic information, the accident details, and the photos of the first accident scene are input into the multimodal large model to determine the rationality of the damage and obtain the authenticity analysis results. The accident details and the photos of the first accident scene are input into the multimodal large model to obtain the cause of the accident; The accident details and the photos of the first accident scene are input into the multimodal large model to obtain the location of the damage; The accident details, the location of the damage, the photos of the accident scene, and business experience are input into a multimodal large model to obtain the estimated loss amount; the business experience is obtained by statistically analyzing the correspondence between parts, damage types, and compensation amounts in historical data.
9. A computer device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the vehicle insurance survey method as described in any one of claims 1-4.
10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed on a terminal device, cause the terminal device to perform the vehicle insurance survey method as described in any one of claims 1-4.