Patent Application
20240403837
The present invention relates to automobile information technology and automobile loss assessment technology, and more particularly to a vehicle accident loss assessment method and apparatus.
In application fields related to vehicle loss determination, such as insurance claims, insurance assessment, traffic accident loss assessment, and vehicle maintenance, how to improve work efficiency, reduce labor costs, improve service levels, and win customer satisfaction through internet technology, IT technology, and/or artificial intelligence technology has become an urgent demand shared by the industry and market. In the process of assessing traffic accident losses, due to factors such as site, equipment, and/or time urgency, it is necessary to conduct a rapid assessment of vehicles with severe losses without dismantling and inspection. In the process of insurance claims and loss assessment, for vehicles with severe losses, it is first necessary to evaluate whether the vehicle has maintenance value, and also necessary to conduct a rapid assessment without dismantling and inspection, so as to determine whether the loss is assessed by normal maintenance or by presumed total loss.
At present, the pain points for insurance companies in assessing the loss of vehicles with severe losses are:
firstly, the time period for assessing the loss by dismantling inspection is long;
secondly, assessing the loss by dismantling inspection is generally carried out in the maintenance enterprise, and even if the insurance company ultimately determines to assess the loss by presumed total loss and conduct a residual value auction of the vehicle, it is obstructed by the maintenance enterprise in every way; and
thirdly, unnecessary vehicle dismantling and inspection costs are increased.
For accident vehicles with severe losses and loss assessment by normal maintenance, the number of parts loss items is mostly in the tens or even hundreds. Manually creating a loss assessment sheet itself is a time-consuming task, and it often takes more than one hour. Rapid assessment, minor correction, and completing the creation of a loss assessment sheet in a few minutes will undoubtedly greatly improve work efficiency and reduce labor costs.
The present invention innovatively provides a vehicle accident loss assessment method and apparatus, which breaks through the traditional operation mode and concept of loss assessment, models the damage form of a vehicle collision, mines and analyzes a large amount of historical data, and realizes a correlation algorithm between the vehicle collision damage models and loss details.
To achieve the above technical objectives, in one aspect, the present invention discloses a vehicle accident loss assessment method. The vehicle accident loss assessment method includes: acquiring vehicle information of a vehicle and repair shop information; virtualizing the vehicle into a corresponding graph in a three-dimensional coordinate system according to the vehicle information, determining a location of a collision portion of the vehicle based on the graph, and determining an accident type; determining core parts involved in the collision portion and a damage forms of the core parts according to the vehicle information, the location of the collision portion in the three-dimensional coordinate system, and the accident type; calculating correlated damaged parts according to the core parts involved in the collision portion and the damage forms of the core parts, so as to obtain a list of damaged parts; and generating a loss report by the list of damaged parts in combination with the repair shop information, wherein the loss report includes a loss assessment value.
Further, for the vehicle accident loss assessment method, the vehicle information includes vehicle brand and configuration vehicle model information.
Further, for the vehicle accident loss assessment method, calculating correlated damaged parts includes: determining a collision damage type according to the core parts involved in the collision portion and the damage forms of the core parts; selecting a corresponding correlation density model according to the determined collision damage type; and calculating the correlated damaged parts by using the selected correlation density model.
Further, for the vehicle accident loss assessment method, the collision damage type has a first correspondence with the core parts and the damage forms of the core parts, the collision damage type has a second correspondence with the correlation density model, and a generation process of the first correspondence and the second correspondence includes: collecting historical case data and classifying vehicle models according to the vehicle information in the historical case data; virtualizing each car in a historical case into a corresponding graph in a three-dimensional coordinate system, and determining a location of a collision portion of the vehicle based on the graph; determining a name of a car collision portion, a height of the car collision on a car body, and a damage degree of the car collision according to an installation location of each part in the three-dimensional coordinate system in vehicles with different brands and different configuration vehicle models, and according to the location of the collision portion of the vehicle and data of the damage degree of the car collision in the historical case data; classifying the collision damage type according to the car collision portion, the height of the car collision on the car body, and the damage degree of the car collision; for different collision damage types of vehicles with different car shapes, separately determining corresponding core parts and damage forms of the core parts, and saving the collision damage types in correspondence with the core parts and the damage forms of the core parts; and for different collision damage types of vehicles with different car shapes, separately establishing corresponding correlation density models, and saving the collision damage types in correspondence with the correlation density models.
Further, for the vehicle accident loss assessment method, and for different collision damage types of vehicles with different car shapes, separately determining corresponding core parts and damage forms of the core parts includes: taking a historical accident vehicle loss assessment record in the historical case data as sample data to perform a big data mining analysis, and classifying the sample data according to a loss amount; and analyzing a probability of various part damages occurring in various vehicle damage types involved in the sample data of different amount segments, so as to determine the corresponding core parts and the damage forms of the core parts.
Further, for the vehicle accident loss assessment method, and for different collision damage types of vehicles with different car shapes, separately establishing corresponding correlation density models includes: taking a historical accident vehicle loss assessment record in the historical case data as sample data to perform a big data mining analysis; and analyzing a probability of damages occurring between parts, so as to generate the correlation density models.
Further, for the vehicle accident loss assessment method, generating a loss report by the list of damaged parts in combination with the repair shop information, wherein the loss report includes a loss assessment value, includes: generating a preview loss report by the list of damaged parts in combination with the repair shop information; performing a deviation correction on a list of parts in the preview loss report; and generating the loss report according to the list of parts after the deviation correction in combination with the repair shop information, wherein the loss report includes the loss assessment value.
To achieve the above technical objectives, in another aspect, the present invention discloses a vehicle accident loss assessment apparatus. The vehicle accident loss assessment apparatus includes: an information acquisition unit, configured to acquire vehicle information of a vehicle and repair shop information; a collision location and accident type determination unit, configured to virtualize the vehicle into a corresponding graph in a three-dimensional coordinate system according to the vehicle information, determine a location of a collision portion of the vehicle based on the graph, and determine an accident type; a core part damage definition unit, configured to determine a core parts involved in the collision portion and a damage forms of the core parts, according to the vehicle information, the location of the collision portion in the three-dimensional coordinate system, and the accident type; an correlated damaged part calculation unit, configured to calculate correlated damaged parts according to the core parts involved in the collision portion and the damage forms of the core parts, so as to obtain a list of damaged parts; and a loss report generation unit, configured to generate a loss report by the list of damaged parts in combination with the repair shop information, wherein the loss report includes a loss assessment value.
To achieve the above technical objectives, in yet another aspect, the present invention discloses a computing device. The computing device includes: one or more processors, and a memory coupled to the one or more processors, the memory storing instructions that, when executed by the one or more processors, cause the one or more processors to perform the above method.
To achieve the above technical objectives, in yet another aspect, the present invention discloses a machine-readable storage medium. The machine-readable storage medium stores executable instructions that, when executed, cause the machine to perform the above method.
The beneficial effects of the present invention are:
The vehicle accident loss assessment method and device provided by embodiments of the present invention is an efficient, intelligent, and convenient vehicle loss assessment model. It breaks the limitation in the current car industry that accident vehicles with severe losses require dismantling inspection for loss assessment, and can determine the loss boundary without dismantling inspection, which is of great significance for insurance companies to develop reasonable claim plans and thereby save claim costs. It breaks the limitation in the current car industry that accident vehicles with severe losses relay too much on employee skills for loss assessment, reduces employee technical requirements, and greatly improves the efficiency of loss assessment. For on-site insurance, the rapid and accurate estimation provides a scientific guarantee for the calculation of outstanding claims reserve of insurance companies, and completely changes the past brainstorming estimation method. In the assessment of road traffic accident losses, the rapid assessment of accident vehicles has profound significance and value for the rapid handling of traffic accidents and/or accident disputes.
In the figures,
Detailed explanations and descriptions to the vehicle accident loss assessment method and apparatus provided by the present invention are made below in conjunction with the accompanying drawings of the specification.
As shown in
As an optional embodiment, the vehicle brand and configuration vehicle model information may be acquired by Optical Character Recognition (OCR) after a user enters a Vehicle Identification Number (VIN) image of a vehicle or by manually entering the Vehicle Identification Number information of a vehicle by a user. The repair shop information such as the name of the repair shop may be acquired. Since the repair prices of the same car parts may be different for different types of repair shops, the type of a repair shop can be judged according to the name of the repair shop. For example, the type of the repair shop may include a 4S shop and a comprehensive repair shop, so as to determine the repair prices of the parts according to the type of the repair shop. The vehicle information may also include license plate number information. When a loss report needs to be sent to an insurance company for claim use, the vehicle information needs to include the license plate number information, and the license plate number information can also be manually entered by the user. If the loss report does not need to be sent to the insurance company, and it is only for a personal inquiry, the license plate number information can also be left blank. As an alternative mode, a case identification code can also be entered or generated.
As a specific example, the basic information of a loss assessment vehicle can be generated through the process of creating an estimate sheet. The user can use a mobile device such as a cell phone to access the APP homepage or use a desktop computer to access the webpage homepage. After the license plate number information is entered, the Vehicle Identification Number information is acquired by OCR recognition after a user enters a Vehicle Identification Number (VIN) image of a vehicle or by manually entering the Vehicle Identification Number information of a vehicle. The vehicle model is determined according to the Vehicle Identification Number information, and then the repair shop information is selected and improved, thereby creating an estimate sheet. An example of the estimate sheet is shown in
In Step S120, the vehicle is virtualized into a corresponding graph in a three-dimensional coordinate system according to the vehicle information, a location of a collision portion of the vehicle is determined based on the graph, and an accident type is determined.
The same type of car brand model can be virtualized into a cuboid with a corresponding size in a three-dimensional coordinate system. The user can draw a circle on the virtual cuboid to describe a collision portion of the vehicle and select an accident type.
Example 1: taking the right rear side collision as an example,
Example 2:
Example 3:
In Step S130, core parts involved in the collision portion and the damage forms of the core parts are determined according to the vehicle information, the location of the collision portion in the three-dimensional coordinate system, and the accident type.
Among them, the core parts of the vehicle can be preset. For example, structural parts of the vehicle, parts that act as collision buffers, and/or parts having high loss amounts after damage can be preset as the core parts. The loss amount of parts is related to the price value of the parts. For example, parts with high loss amounts is the parts with the value above a predetermined threshold or the parts with a ratio of the value of the part to the total vehicle value greater than a threshold (such as 10%). For example, the front longitudinal beam assembly (left) is a structural component that has the greatest collision bearing force; the front bumper frame is a part that acts as a collision buffer; the value of the high-voltage battery is 30% of the total vehicle value, exceeding the threshold of 10%. Therefore, the front bumper frame, the front longitudinal beam assembly, and the high-voltage battery can be predefined as core parts. The core parts of the vehicle are distributed in the corresponding vehicle portions. Based on vehicle information, the relationship between the core parts and the vehicle portions can be determined. Therefore, according to the collision portion described in the circle, the pre-set core parts of the vehicle and their relationship with the vehicle portions, the core parts corresponding to the collision portion can be determined. Also, a correlation between the core part and the accident type may be preset, and thus the core part can be determined from the collision portion and the accident type.
Specifically, the loading positions of car parts are first coordinated. Based on the actual loading position of the parts, a coordinate system of the loading parts is established, and a three-dimensional coordinate is given to the loading position of each part.
As a specific example, after the operation that the user circles and ticks the accident type, the program will pop up a corresponding list of core parts according to the actual loaded parts of the specific vehicle, and the user is required to judge the type of damage: whether to replace the damaged parts or repair them with sheet metal. Among them, the core parts can be parts with a damage form that can be seen or can be visually inspected on the appearance of the vehicle, or parts whose damaged form can be judged without being disassembled and inspected. The number of core parts can be between 3-12, and can vary depending on the number of vehicle collision portions and the vehicle configuration.
In Step S140, correlated damaged parts are calculated according to the core parts involved in the collision portion and the damage forms of the core parts, so as to obtain a list of damaged parts.
Calculating correlated damaged parts in Step S140 may include the following steps: determining a collision damage type according to the core parts involved in the collision portion and the damage forms of the core parts; selecting a corresponding correlation density model according to a determined collision damage type; and calculating the correlated damaged parts by using the selected correlation density model. The names of all damaged parts obtained are collected, and the redundant parts are filtered out based on the actual accident vehicle loading parts to generate a list of damaged parts.
In Step S150, a loss report is generated by the list of damaged parts in combination with the repair shop information, wherein the loss report includes a loss assessment value.
Generating a loss report by the list of damaged parts in combination with the repair shop information in Step S150 may include the following steps: generating a preview loss report by the list of damaged parts in combination with the repair shop information; performing a deviation correction on a list of parts in the preview loss report; and generating the loss report according to the list of parts after the deviation correction in combination with the repair shop information, wherein the loss report includes the loss assessment value.
As a specific example, as shown in
The reason for the deviation is that there is an error in the damage form of “replace” or “repair” for a certain number of parts, and the user needs to perform a deviation correction again for an accurate and complete assessment. As an optional embodiment, the deviation correction can be performed by the user to delete and/or add the core parts in the list of damaged core parts in the preview loss report, and/or modify the damage forms of the core parts, and so on.
As an optional embodiment, in Step S140 above, the collision damage type may have a first correspondence with the core parts involved in the collision portion and the damage forms of the core parts, and the collision damage type may have a second correspondence with the correlation density model. As shown in
In Step S1410, historical case data are collected and vehicle models are classified according to the vehicle information in the historical case data.
The body shape of a passenger car may be mainly divided into six types: three-box four-door, three-box two-door, two-box five-door, SUV, MPV, and van. The specific shape form of each vehicle can be acquired by parsing the Vehicle Identification Number (VIN) of the vehicle.
In Step S1420, each car in a historical case is virtualized into a corresponding graph in a three-dimensional coordinate system, and a location of a collision portion of the vehicle is determined based on the graph.
Specifically, it is necessary to create three-dimensional coordinate data of the car collision portion and area in advance. As shown in
In Step S1430, a name of a car collision portion, a height of the car collision on a carbody, and a damage degree of the car collision are determined according to an installation location of each part in the three-dimensional coordinate system in vehicles with different brands and different configuration vehicle models, and according to the location of the collision portion of the vehicle and data of the damage degree of the car collision in the historical case data.
Specifically, it is necessary to create three-dimensional coordinate gyroscope data for the installation locations of car parts in advance. In a car collision accident, the part names in the car part database, such as the TOP2000 part database, can cover 99.90% of car parts.
Which parts in the TOP2000 part database are actually installed in each vehicle can be acquired by parsing the Vehicle Identification Number (VIN) of the vehicle. In the TOP2000 part database, 80% of the parts are respectively installed in the same position among six body shapes of three-box four-door, three-box two-door, two-box five-door, SUV, MPV, and van, and the other 20% of the parts vary depending on the car brand and the configuration vehicle model.
The six body shapes can be divided into two categories, and two three-dimensional part coordinate systems are respectively established as follows:
A first part coordinate system: the part coordinate system for vehicles with body shapes other than the “three-box two doors” is OXYZ=0 (9,12,3), and the origin O=(0,0,0) is the left front lower corner point of the vehicle; the X-axis represents the left and right direction of the vehicle, with a maximum scale of 12; the Y-axis represents the front and rear direction of the vehicle, with a maximum scale of 12; and the Z-axis represents the up and down direction of the vehicle, with a maximum scale of 3.
A second part coordinate system: the part coordinate system for a vehicle with a body shape of “three-box two doors” is OXYZ=0 (9,10,3), and the origin O=(0,0,0) is the left front lower corner point of the vehicle; the X-axis represents the left and right direction of the vehicle, with a maximum scale of 9; the Y-axis represents the front and rear direction of the vehicle, with a maximum scale of 10; and the Z-axis represents the up and down direction of the vehicle, with a maximum scale of 3.
As a specific example, more than 100 passenger car brands, over 200,000 different configuration vehicle models, and 2000 part names can be used to create three-dimensional coordinate gyroscope data of car part installation locations in a computer mode based on the actual part installation locations.
In Step S1440, the collision damage type is classified according to the car collision portion, the height of the car collision on the carbody, and the damage degree of the car collision.
As a specific example, 156 collision damage types can be created according to the car collision portion, the height of the car collision on the car body, and the damage degree of the car collision. As shown in Table 1 below, there are 10 collision damage types in the left front corner and 15 collision damage types in the left front side.
As a specific example, the correspondence between the core parts and the part damage forms and the 156 collision damage types is obtained through a data analysis method. In Step S1450, for different collision damage types of vehicles with different car shapes, the corresponding core parts and the damage forms of the core parts are separately determined, and the collision damage types are saved in correspondence with the core parts and the damage forms of the core parts.
Specifically, Step S1450 may include the following steps: taking a historical accident vehicle loss assessment record in the historical case data as sample data to perform a data analysis, and classifying the sample data according to a loss amount; and analyzing a probability of various part damages occurring in various vehicle damage types involved in the sample data of different amount segments, so as to determine the corresponding core parts and the damage forms of the core parts.
As a specific example, the loss assessment records of the historical accident vehicles in tens millions level are specifically taken as samples. The data analysis method is used to classify the samples according to loss amount segments, and analyze the probability of various part damages occurring in various vehicle damage types involved in the sample data of different amount segments, so as to determine the core key part names and part damage forms of 156 collision damage types. For example, the samples are divided with loss amount segments such as 5,000-10,000 RMB, 10,000-100,000 RMB, 100,000-500,000 RMB. By statistically calculating the sample data, in a high position of the left front side, the probability of part A damage occurring (the number of part A damage divided by the total number of samples in the 5,000-10,000 segment) is 90%, the probability of part B damage occurring in the 5,000-10,000 segment is 50%, and the probability of part C damage occurring in the same segment is 30%, wherein the parts can be predetermined, for example including structural parts of the vehicle, parts that act as collision buffers, and/or parts having high loss amounts. The part with a probability of damage occurring greater than a predetermined threshold (e.g. 80%) can be determined as the core key part. In this example, in the high position of the left front side and in the 5,000-10,000 segment, the core key part is part A, and the damage form of the part A can be determined from the historical case data. Thus, the core key parts and the part damage forms under each collision portion, each height of the car collision, and each loss amount segment can be determined. Different loss amounts/loss amount segment may correspond to different damage degree of the car collision (such as light, moderate, or heavy collision). Therefore, with the analysis of the loss amount of sample data, it is possible to determine the core key part and part damage forms of the collision damage types, thereby obtaining the correspondence between the core key parts and the part damage forms and the collision damage types.
On the contrary, when the user circles and ticks the core part damage forms, the computer can intelligently calculate the damage type of the accident vehicle.
There is a strong correlation between the collision portion and the accident type of a vehicle and the list of key core parts, which can be obtained through the data analysis method. In Step S1460, for different collision damage types of vehicles with different car shapes, the corresponding correlation density models are separately established, and the collision damage types are saved in correspondence with the correlation density models.
Specifically, Step S1460 may include the following steps: taking a historical accident vehicle loss assessment record in the historical case data as sample data to perform data analysis; and analyzing a probability of the damages occurring between parts, so as to generate the correlation density models. As a specific example, the loss assessment records of the historical accident vehicles in tens millions level are also taken as samples to analyze the correlation density between part damages for different collision damage types. A neural density algorithm can be used to analyze the correlation density between part damages. For example, when a part A undergoes a replacement damage form, what is the density of a part B undergoing a replacement damage form. For another example, when both part A and part B undergo the replacement damage form, what is the density of a part C undergoing a replacement damage form.
For Example, the neural density algorithm includes a calculation of a probability of occurrence of damage between the parts (that is, a damage also occurs to another part when a damage occurs to one part) based on the historical data. Bayes' theorem may be used for the calculation, expressed s Confidence
where Support (A) represents the probability that a damage occurs to part A in the historical data, namely, a count number of damage of part A divided by all count number of the historical data (e.g., A_count/all count: 409296/5861998 in Table 2); and Support (A U B) represents the probability that a damage occurs to part A and part B at the same time in the historical data, namely, a count number of damage of part A and part B divided by all count number of the historical data (e.g., A_B_count/all count: 362630/5861998 in Table 2). When the calculated probability (Confidence) that a damage occurs to part A and part B at the same time is greater than a predetermined threshold, it may be judged that there is a correlation between part A and part B. Thus, the correlation between the parts can be obtained based on the calculated probability of the damages occurring between parts, to generate correlation density models.
Front bumper frame
Front bumper crust
Front bumper frame
Center Grid
Front bumper frame
Engine hood
Front bumper frame
Front bumper
Front bumper frame
Headlight assembly
left
replacement
Front bumper frame
Headlight assembly
right
replacement
Front bumper frame
Front fender
left
Front bumper frame
or frame
A
B
A
B
A
B
indicates data missing or illegible when filed
Through the above table 2, in the 5861998 historical vehicle accident damage cases, it is calculated that when the damage item of {Front bumper frame replacement}, i.e., the damage event of part A occurs, the probability that the damage item of {Front bumper cover replacement}, i.e., the damage event of part B occurs, is 0.885985 (“Confidence”), which is greater than a predetermined threshold (e.g., 0.5). Thus, it can be judged that there is a correlation between {Front bumper frame replacement} and {Front bumper cover replacement}.
In addition, the correlation between the parts may be determined by further calculating “Lift” and “Conviction” in addition to “Confidence”, to further improve the accuracy of the correlation density model. “Lift” is to measure the independence of damages between the two parts A, B, Lift
where Support (B) represents the probability that a damage occurs to part B in the historical data, namely, a count number of damage of part B divided by all count number of the historical data (e.g., B_count/all count: 1538505/5861998 in Table 2); and “Conviction” is to measure the probability of the rule prediction error, and represents the probability that event A occurs while event B does not occur.
The strength of the correlation between the vehicle parts may be determined by calculating the “Confidence”, the “Lift” and the “Conviction”. For example, the strong correlation between the vehicle parts means that when one vehicle part is damaged and thus becomes a damaged part, a probability that another vehicle part also becomes a damaged part accordingly is higher than a predetermined threshold (e.g., 90%). In more detail, when “Confidence” is higher than a predetermined threshold, it is confirmed whether “Support”, “Lift” and “Conviction” are also higher than their respective predetermined thresholds, respectively, and if the confirmation result is positive, then it is considered that there is a strong correlation between the vehicle parts.
For example, with referring to Table 2, when creating a vehicle parts damage correlation density model, the record with a total of 5861998 historical vehicle accident cases is taken as a sample to calculate the probability that a damage item occurs to part B when a damage item occurs to part A. Here, it is assumed that A={Front bumper frame replacement}, which represents a damage event of part A; B={Engine hood painting}, {Front bumper liner replacement}, {Front bumper crust replacement}, {Headlight assembly (right) replacement}, {Headlight assembly (left) replacement}, {Front fender (left) painting}, {Radiator frame replacement}, or {Center Grid Replacement}, etc., which represent damage events of part B. In order to ensure the strong correlation between the damage event of part A and the damage events of part B, the thresholds are set: for example, Confidence (A->B)≥0.5; Lift (a->b)≥0.5; and Conviction (a->b)>1. At this point, based on the sample data of these historical vehicle accident cases, as shown in the Table 2, and according to the above-mentioned formulas, the “Support”, “Confidence”, “Lift” and “Conviction” in each corresponding situation can be calculated.
Thus, in the 5861998 historical vehicle accident damage cases, it is calculated that when the damage item of {Front bumper frame replacement}, i.e., the damage event of part A occurs, the probability that the damage item of {Front bumper cover replacement}, i.e., the damage event of part B occurs, is 0.885985 (Confidence), which belongs to the strong correlation event. Thus, it can be judged that there is a strong correlation between {Front bumper frame replacement} and {Front bumper cover replacement}.
By analogy, the part damage forms that are strongly correlated to the parts in the TOP2000 part database are find out.
Through the correlation density model, it is possible to obtain the correlated damage parts in addition to the core damage parts, and thus it is possible to generate a more comprehensive list of damaged parts, thereby improving the accuracy of the loss report.
The information acquisition unit 1110 is configured to acquire vehicle information of a vehicle and repair shop information. The operation of the information acquisition unit 1110 can refer to the operation of Step S110 described above with reference to
The collision location and accident type determination unit 1120 is configured to virtualize the vehicle into a corresponding graph in a three-dimensional coordinate system according to the vehicle information, determine a location of a collision portion of the vehicle based on the graph, and determine an accident type. The operation of the collision location and accident type determination unit 1120 can refer to the operation of Step S120 described above with reference to
The core part damage definition unit 1130 is configured to determine a core part involved in the collision portion and a damage form of the core part, according to the vehicle information, the location of the collision portion in the three-dimensional coordinate system, and the accident type. The operation of the core part damage definition unit 1130 can refer to the operation of Step S130 described above with reference to
The correlated damaged part calculation unit 1140 is configured to calculate correlated damaged parts according to the core parts involved in the collision portion and the damage forms of the core parts, so as to obtain a list of damaged parts. The operation of the correlated damaged part calculation unit 1140 can refer to the operation of Step S140 described above with reference to
The loss report generation unit 1150 is configured to generate a loss report by the list of damaged parts in combination with the repair shop information, wherein the loss report includes a loss assessment value. The operation of the loss report generation unit 1150 can refer to the operation of Step S150 described above with reference to
The vehicle accident loss assessment method and apparatus provided by embodiments of the present invention are characterized by coordinating the car part loading position, virtualizing the same type of car brand models into a cuboid, and then establishing a coordinate system of the loading parts based on the actual loading position of the parts, and giving each part a three-dimensional coordinate of the loading position, and at the same time, mining and analyzing historical accident vehicle assessment records, and using a neural density algorithm to calculate the probability of damage occurring between vehicle parts and the correlation between the vehicle part damage and the damage degree of vehicle collision. The mathematical model of the relationship between the vehicle accident collision portion, the damage degree and the part damage is established, so as to quickly and intelligently evaluate the accident vehicle loss report. Artificial Intelligence (AI) is used to model the vehicle collision damage form, and a large amount of historical data are used for mining and analyzing, thereby realizing the correlation algorithm between the vehicle collision damage model and the loss detail, which can be conveniently operated and used by users at the mobile phone terminal and the webpage terminal, for quick assessment.
As shown in
In one embodiment, the memory 1220 stores computer executable instructions which, when executed, cause the at least one processor 1210 to execute: acquiring vehicle information of a vehicle and repair shop information; virtualizing the vehicle into a corresponding graph in a three-dimensional coordinate system according to the vehicle information, determining a location of a collision portion of the vehicle based on the graph, and determining an accident type; determining a core part involved in the collision portion and a damage form of the core part according to the vehicle information, the location of the collision portion in the three-dimensional coordinate system, and the accident type; calculating correlated damaged parts according to the core part involved in the collision portion and the damage form of the core part, so as to obtain a list of damaged parts; and generating a loss report by the list of damaged parts in combination with the repair shop information, wherein the loss report includes a loss assessment value.
It should be understood that the computer executable instructions stored in the memory 1220, when executed, cause at least one processor 1210 to perform various operations and functions described above in conjunction with
In this disclosure, the computing device 1200 may include, but is not limited to: a personal computer, a server computer, a workstation, a desktop computer, a laptop computer, a notebook computer, a mobile computing device, a smartphone, a tablet, a cellular phone, a personal digital assistant (PDA), a handheld device, a messaging device, a wearable computing device, a consumer electronic device, and the like.
According to one embodiment, a program product such as a non-temporary machine-readable medium is provided. The non-temporary machine-readable medium may have instructions (i.e., the above elements implemented in software form) that, when executed by the machine, cause the machine to perform various operations and functions described above in conjunction with
Specifically, a system or device equipped with a readable storage medium can be provided, in which a software program code that realizes the functions of any of the above embodiments is stored, and the computer or processor of the system or device is caused to read and execute instructions stored in the readable storage medium.
In this case, the program code per se read from the readable medium can realize the functions of any of the above embodiments, and thus the machine readable code and the readable storage medium storing the machine readable code constitute a part of the present invention.
Embodiments of the readable storage medium include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, and DVD-RW), a magnetic tape, a non-volatile memory card, and ROM. Alternatively, the program code can be downloaded from the server computer or the cloud through the communication network.
The above described are only embodiments of the present invention and do not limit the protection scope of the claims of the present invention. Any equivalent structure or equivalent process transformation made by using the description and accompanying drawings of the present invention, or directly or indirectly applied to other related technical fields, are similarly included in the protection scope of the claims of the present invention.
| Number | Date | Country | Kind |
|---|---|---|---|
| 202110836463.1 | Jul 2021 | CN | national |
| Number | Date | Country | |
|---|---|---|---|
| Parent | PCT/CN2022/088381 | Apr 2022 | WO |
| Child | 18224323 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 18224323 | Jul 2023 | US |
| Child | 18800124 | US |
