Data processing method and device, equipment and storage medium

By analyzing accident scene photos and user supplementary information, combining vehicle maintenance and insurance information, identifying accident types and responsibilities, the problems of low efficiency in handling traffic accidents and uncertainty in determining liability are solved, and a rapid and accurate accident handling and claim processing process is achieved. .

CN119048805APending Publication Date: 2024-11-29CHONGQING SELIS PHOENIX INTELLIGENT INNOVATION TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411051626.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-01
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

The existing traffic accident handling methods are inefficient in handling traffic accidents, which are prone to major traffic jams and secondary accidents. The judgment of responsibility depends on manual judgment, which has subjectivity and uncertainty.

Method used

By receiving and analyzing accident scene photos, using pre-trained accident type judgment model to identify accident type and responsibility allocation results, and through actual proportion screening and user supplementary information correction, the final accident type and responsibility allocation results are generated, and combined with vehicle maintenance. Database and insurance information, providing reasonable compensation advice.

Benefits of technology

It realizes rapid recording and analysis of traffic accidents, improves processing efficiency, reduces safety hazards, reduces uncertainty in subjective judgments, simplifies the claim settlement process, and improves the claim success rate.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119048805A_ABST
    Figure CN119048805A_ABST
Patent Text Reader

Abstract

The embodiment of the invention provides a data processing method and device, equipment and a storage medium, and the method comprises the steps: receiving a plurality of current accident scene pictures uploaded by a user, recognizing the first key feature information, inputting the first key feature information into a pre-trained accident type judgment model, outputting a first accident type prediction result and a first responsibility allocation result about the current accident based on each current accident scene picture, and determining a correct second accident type prediction result and a correct second responsibility allocation result according to the proportion of each accident type in the first accident type prediction result, when the maximum actual proportion is smaller than a first preset value, a first supplementary prompt is sent, key information input by a user for the first supplementary prompt is received, and a third accident type prediction result and a third responsibility allocation result about the accident type of the current accident are generated. And the reasonability of the conclusion is ensured.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of traffic management, and in particular to a data processing method, device, equipment and storage medium. Background Art

[0002] With the development of society, vehicles are becoming more and more popular in people's daily lives, and the increasing number of vehicles will also lead to traffic congestion on daily commuting roads. Big data shows that in addition to objective factors such as excessive traffic flow and inadequate road infrastructure, the problem of traffic congestion during peak hours is often caused by subjective factors such as a small number of drivers engaging in dangerous overtaking, changing lanes at will, and other uncivilized behaviors in a hurry, which easily leads to minor traffic accidents such as scratches and rear-end collisions.

[0003] After an accident occurs, it is usually necessary to wait for the traffic police to arrive at the scene to determine the responsibility. This method has the following disadvantages: first, it is difficult for the traffic police to quickly arrive at the scene of the accident for immediate processing due to road congestion. Not only is the processing efficiency low, but it may even cause a large traffic jam caused by a small accident, or even cause a secondary traffic accident, posing a major traffic safety hazard. Secondly, the traffic police arrive at the scene after the accident occurs. If the accident scene is not preserved intact, it is impossible to objectively and completely understand the course of the accident at the scene. Moreover, the determination of accident responsibility mainly relies on manual judgment, which may be subjective and uncertain. Summary of the invention

[0004] The purpose of the embodiments of the present invention is to provide a data processing method, device, equipment and storage medium to solve the problems that the existing data processing is not only inefficient, but may even cause a large traffic jam caused by a small accident, or even cause a secondary traffic accident, which poses a great traffic safety hazard. In addition, the accident responsibility determination mainly relies on manual judgment, which may be subjective and uncertain. The specific technical solution is as follows:

[0005] In a first aspect of the present invention, a data processing method is provided, the method comprising:

[0006] Receiving a number of current accident scene photos uploaded by a user, and identifying first key feature information in the number of current accident scene photos;

[0007] Inputting the first key feature information into a pre-trained accident type determination model, and outputting a first accident type prediction result and a first responsibility allocation result for the current accident based on each of the current accident scene photos;

[0008] Obtaining the types of accidents included in the first accident type prediction result and the actual proportion of each of the accident types in a number of current accident scene photos;

[0009] Filter out the largest actual proportion from the actual proportions and compare it with the first preset value;

[0010] If the maximum actual proportion is greater than the first preset value, determining the correct second accident type prediction result and the second responsibility allocation result from the first accident type prediction result and the first responsibility allocation result according to the maximum actual proportion;

[0011] If the maximum actual proportion is less than the first preset value, sending a first supplementary prompt;

[0012] Receive key information input by the user in response to the first supplementary prompt, and generate a third accident type prediction result and a third responsibility allocation result regarding the accident type of the current accident.

[0013] Optionally, before receiving a plurality of current accident scene photos uploaded by a user and identifying the first key feature information in the plurality of current accident scene photos, the method further includes:

[0014] Collect at least one historical accident scene photo;

[0015] Marking each of the historical accident scene photos with the corresponding accident type and responsibility allocation result and obtaining key feature information of each of the historical accident scene photos, the key feature information including road information and vehicle information, and the accident types including rear-end collisions, overtaking and lane change accidents, reversing and rolling accidents, intersection and roundabout accidents, and dedicated lane accidents;

[0016] Establishing a correlation relationship between the accident type, responsibility allocation results and key feature information of each of the historical accident scene photos;

[0017] The accident type determination model is trained through at least one of the association relationships, wherein key feature information in each association relationship is used as input information, and the accident type and responsibility allocation result are used as corresponding output information.

[0018] Optionally, before inputting the first key feature information into a pre-trained accident type determination model and outputting a first accident type prediction result and a first responsibility allocation result about the current accident based on each current accident scene photo, the method further includes:

[0019] If it is detected that the amount of the first key feature information identified is less than the second preset value, sending a second supplementary prompt;

[0020] In response to the user's supplementary input for the second supplementary prompt, the first key feature information is updated.

[0021] Optionally, after receiving the key information input by the user in response to the first supplementary prompt and generating a third accident type prediction result and a third responsibility allocation result regarding the accident type of the current accident, the method further includes:

[0022] Collect original pictures of different parts of different models of vehicles;

[0023] Comparing the original image with the current accident scene photo to determine second key feature information about the current vehicle in the current accident scene photo, the second key feature information including the current vehicle model, the damaged part of the current vehicle, and the damage type of the current damaged part;

[0024] The first repair cost of each damaged part of the current vehicle is determined by searching the preset vehicle repair database through the second key feature information, wherein different repair costs correspond to different types of damage to the same damaged part of the same vehicle model, and different repair costs correspond to the same damaged part of different vehicle models and the same type of damage to the same vehicle model;

[0025] Determine the target number of damaged parts of the current vehicle through the second key feature information;

[0026] The target number of the first maintenance costs is accumulated to generate the second maintenance cost of the current vehicle.

[0027] Optionally, after receiving the key information input by the user in response to the first supplementary prompt and generating a third accident type prediction result and a third responsibility allocation result regarding the accident type of the current accident, the method further includes:

[0028] Get the user's insurance information for the current vehicle;

[0029] Determine the target conditions for the effectiveness of the insurance policy through the insurance information;

[0030] Determining whether the third accident type prediction result and the third responsibility allocation result meet the target condition;

[0031] If satisfied, then determining the target policy information corresponding to the target condition;

[0032] The target compensation amount is determined by the target policy information.

[0033] Optionally, after determining the target compensation amount obtained through the target insurance policy information, the method further includes:

[0034] Retrieving the third accident type prediction result, the third responsibility allocation result and the target repair cost for the current vehicle from a preset vehicle repair database;

[0035] Determine the additional amount of premium for reinsurance through the search results, wherein the additional amount of premium is the additional cost of reinsurance over the current insurance for the same type of insurance;

[0036] comparing the target compensation amount and the premium increase amount;

[0037] Generate reasonable compensation recommendations by comparing the results.

[0038] In a second aspect of the present invention, there is also provided a data processing device, comprising:

[0039] A first recognition module is used to receive a number of current accident scene photos uploaded by a user, and recognize first key feature information in the number of current accident scene photos;

[0040] A first output module, configured to input the first key feature information into a pre-trained accident type determination model, and output a first accident type prediction result and a first responsibility allocation result for the current accident based on each of the current accident scene photos;

[0041] A first acquisition module, used to acquire the types of accidents included in the first accident type prediction result and the actual proportion of each of the accident types in a number of current accident scene photos;

[0042] A first screening module, used to screen out the largest actual proportion from the actual proportions and compare it with a first preset value;

[0043] A first determination module, configured to determine a correct second accident type prediction result and a second responsibility allocation result from the first accident type prediction result and the first responsibility allocation result by using the maximum actual proportion if the maximum actual proportion is greater than a first preset value;

[0044] A first sending module, configured to send a first supplementary prompt if the maximum actual proportion is less than a first preset value;

[0045] The first generating module is used to receive key information input by the user in response to the first supplementary prompt, and generate a third accident type prediction result and a third responsibility allocation result regarding the accident type of the current accident.

[0046] In a third aspect of the implementation of the present invention, there is also provided a communication device, comprising: a transceiver, a memory, a processor, and a program stored in the memory and executable on the processor;

[0047] The processor is used to read the program in the memory to implement and execute any of the above-mentioned data processing methods.

[0048] In a fourth aspect of the implementation of the present invention, a computer-readable storage medium is further provided, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute any of the above-mentioned data processing methods.

[0049] A fifth aspect of an embodiment of the present application provides a vehicle, comprising: the above-mentioned data processing device.

[0050] The data processing method provided by the embodiment of the present invention receives a number of current accident scene photos uploaded by a user, and identifies the first key feature information in the number of current accident scene photos, so as to facilitate the subsequent identification of the accident type through this key feature information, input the first key feature information into a pre-trained accident type determination model, and output a first accident type prediction result and a first responsibility allocation result about the current accident based on each of the current accident scene photos; obtains the accident type types contained in the first accident type prediction result and the actual proportion of each of the accident type types in the number of current accident scene photos; selects the largest actual proportion from the actual proportions and compares it with a first preset value; if the largest actual proportion is greater than the first preset value, determines the correct second accident type prediction result and the second responsibility allocation result from the first accident type prediction result and the first responsibility allocation result through the largest actual proportion, and takes the largest actual proportion to represent the most likely recognition result. If the probability of this result is greater than the first preset value, the representativeness of this result is further determined. If the largest actual proportion is less than the first preset value, a first supplementary prompt is sent, and the user's response to the The key information input by the first supplementary prompt is used to generate a third accident type prediction result and a third responsibility allocation result about the accident type of the current accident. Through the user's supplementary input, the output first accident type prediction result is corrected, so that the obtained result is closer to the actual situation of the accident and the responsibility identification is more accurate. The embodiment of the present invention recognizes the accident type to which the accident scene photo belongs and determines the responsibility allocation result through a pre-trained accident type determination model, thereby realizing rapid recording and analysis of traffic accidents, improving the handling efficiency of accidents, and avoiding the occurrence of greater safety hazards. By supplementing key information, the rationality of the conclusion is guaranteed and the uncertainty of subjective judgment is reduced. In addition, the present application also determines the vehicle's maintenance costs and obtains the vehicle's insurance status by identifying the components of the damaged parts of the vehicle in combination with a pre-set vehicle maintenance database. Based on the analysis of the insurance status and maintenance costs, the car owner can clearly and quickly obtain accident handling suggestions and insurance claim details, which not only helps the car owner make more information-supported decisions, but also reduces waiting time and alleviates road congestion. Coupled with direct evidence support and intelligent recommendations, the claim process is simplified and the claim success rate is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for describing the embodiments or the prior art are briefly introduced below.

[0052] Figure 1 A flowchart of a data processing method provided by an embodiment of the present invention;

[0053] Figure 2 A flowchart of another data processing method provided by an embodiment of the present invention;

[0054] Figure 3 A flowchart of another data processing method provided by an embodiment of the present invention;

[0055] Figure 4 is a structural schematic diagram of a data processing device provided by an embodiment of the present invention;

[0056] Figure 5 It is a structural diagram of a communication device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0057] To make the purpose, technical scheme and advantages of the embodiments of the present invention clearer, the embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings. However, it will be appreciated by those skilled in the art that in the embodiments of the present invention, many technical details are proposed in order to enable the reader to better understand the present application. However, even without these technical details and various changes and modifications based on the following embodiments, the technical scheme claimed in the present application can be implemented. The division of the following embodiments is for the convenience of description, and should not constitute any limitation on the specific implementation of the present invention. The various embodiments can be combined and referenced with each other without contradiction.

[0058] Reference Figure 1 , shows a flow chart of steps of a data processing method provided by an embodiment of the present invention, the method may include:

[0059] Step 101, receiving a number of current accident scene photos uploaded by a user, and identifying first key feature information in the number of current accident scene photos.

[0060] In the embodiment of the present invention, after determining that an accident has occurred, the car owner obtains the current accident scene photo through a handheld terminal (such as a mobile phone, a tablet, a phone watch, etc.) and a car camera. There may be several photos of the current accident scene uploaded, and then the car owner uploads the current accident scene photo through a mobile phone APP. After receiving the photo, the cloud server performs preprocessing to help improve the accuracy and efficiency of subsequent image analysis. The following are some common preprocessing steps: resizing: according to the analysis requirements, adjust the photo to a uniform size, cropping: remove irrelevant parts of the photo, focus on the area containing key information, reduce the amount of data processed, and improve the focus of the analysis, enhance contrast: by adjusting the brightness and contrast of the image, the details in the image can be made clearer, which helps the algorithm to better identify and analyze the image content, denoising: using filters or other algorithms to remove noise in the image and improve image quality, color correction: ensure that the color of the image is accurate, histogram equalization: by adjusting the pixel intensity distribution of the image, the contrast of the image is improved, which helps to improve the readability of the image, normalization: convert the pixel value of the image to a specific range, usually 0 to 1 or -1 to 1, which helps the stability and convergence of the algorithm. The cloud server then analyzes the uploaded photos of the current accident scene through the built-in CNN model (accident type determination model). The model includes an image recognition algorithm that can identify the first key feature information in the image. The image recognition algorithm can use the algorithm existing in the current technology, and the present invention will not go into details here. After identifying these features, the accident type determination model will predict the accident type based on these features.

[0061] It should be noted that the accident type determination model needs to be pre-trained. During the training, a large number of accident scene photos containing key feature information are prepared, and then the corresponding accident types are marked, and the corresponding responsibility allocation results are annotated according to the accident type. Among them, it is necessary to ensure that the amount of data collected from the accident scene photos is large enough, and the accident types included are rich enough. Then, based on the large amount of accident type image data marked as above, the model is trained. During the training, the key feature information on the accident scene photos is input according to the one-to-one correspondence of the pictures, and the marked accident type is output. The specific steps include:

[0062] Collect at least one historical accident scene photo;

[0063] Label each historical accident scene photo with the corresponding accident type and responsibility allocation result and obtain the key feature information of each historical accident scene photo. The key feature information includes road information and vehicle information. The accident types include rear-end collision, overtaking and lane change accidents, reversing and sliding accidents, intersection and roundabout accidents and special lane accidents.

[0064] Establish a correlation between the accident type, responsibility allocation results and key feature information of each historical accident scene photo;

[0065] The accident type determination model is trained through at least one association relationship, wherein key feature information in each association relationship is used as input information, and the accident type and responsibility allocation result are used as corresponding output information.

[0066] The trained accident type determination model can adopt the Google net model. The trained accident type determination model can identify key features of accident scene photos and determine the accident type based on the identified key features. The trained accident type determination model is deployed to the car owner's AP P cloud platform is used to predict new data results. The key feature information includes road information and vehicle information, which can also be divided. Specific road information includes lane line position, lane line type, road markings, traffic lights, and road conditions. Vehicle information includes vehicle relative position, vehicle model, and vehicle damaged parts (front, rear, side, wheels, etc.); accident types include rear-end collisions, overtaking and lane changing accidents, reversing and sliding accidents. Intersections and roundabouts and dedicated lane accidents can also be divided. Specific overtaking and lane changing accidents include overtaking accidents that cross the line, lane change accidents, meeting and overtaking accidents, overtaking left turn and U-turn vehicle accidents. Intersections and roundabouts include U-turn failure to give way accidents, vehicles entering or crossing the road failure to give way accidents, vehicles entering the roundabout failure to give way accidents, and vehicles on the right side of the intersection without traffic lights. Dedicated lane accidents include accidents entering dedicated lanes. The responsibility allocation results include full responsibility, primary responsibility, equal responsibility, secondary responsibility, no responsibility and multiple responsibilities.

[0067] Furthermore, when marking the responsibility classification results, the corresponding responsibility allocation results will be determined according to the determination of the accident type. For example, in a rear-end collision, usually, the rear vehicle rear-ends the front vehicle, and the rear vehicle bears the main responsibility or full responsibility because the rear vehicle fails to maintain a safe distance. In a lane change accident, the lane-changing vehicle fails to ensure a safe distance or fails to turn on the turn signal, resulting in the accident. The lane-changing vehicle bears the main responsibility or full responsibility. In a wrong-way accident, the wrong-way vehicle usually bears the main responsibility or full responsibility. In an intersection accident, the vehicle that fails to obey the traffic lights or traffic signs usually bears the main responsibility or full responsibility. At intersections without traffic lights, the vehicle that fails to obey the "give way to the right principle" usually bears the main responsibility or full responsibility. In a reversing accident, the reversing vehicle usually bears the main responsibility or full responsibility for the accident due to failure to observe the surrounding environment or ensure a safe distance when reversing, and the reversing vehicle usually bears the main responsibility or full responsibility, and so on.

[0068] Step 102, input the first key feature information into a pre-trained accident type determination model, and output a first accident type prediction result and a first responsibility allocation result for the current accident based on each current accident scene photo.

[0069] In the embodiment of the present invention, a pre-trained accident type determination model is used to output a first accident type prediction result regarding the accident type based on the first key feature, and a first responsibility allocation result is determined based on the predicted accident type. However, since the user may upload several current accident scene photos, a corresponding first accident type prediction result and a first responsibility allocation result are output based on each current accident scene photo.

[0070] In addition, during the prediction process, if it is detected that the key feature information identified in the uploaded photo is insufficient or missing, the user will be prompted to upload additional on-site photos or to supplement them in text form. The specific steps include:

[0071] If it is detected that the amount of the first key feature information identified is less than the second preset value, sending a second supplementary prompt;

[0072] In response to the user's supplementary input for the second supplementary prompt, the first key feature information is updated.

[0073] Among them, the second preset value is set according to the type of key features set during training. For example, the key features set during the above training include lane line position, lane line type, vehicle relative position, vehicle model, road markings, traffic lights, road conditions and damaged parts of the vehicle. Then the corresponding second preset value can be set to 8. It should be noted that the key features set during training in the embodiment of the present invention can also be supplemented according to needs, and the corresponding second preset value will also change. The present invention does not make specific limitations here. In addition, the second supplementary prompt sent can be a text prompt such as "The key features currently identified lack road condition information, please supplement". When the user supplements, it is supplemented according to the content of the prompt. For some features that cannot be supplemented through pictures, the user needs to enter text. For example, the current road section is far away from the traffic light and it is impossible to obtain photos of the current traffic light, then the user needs to enter text to describe this situation. Through the above prompts, the first key feature information can be completed, and the output results of the subsequent accident type determination model are more reasonable.

[0074] Step 103: Obtain the accident types included in the first accident type prediction result and the actual proportion of each accident type in a number of current accident scene photos.

[0075] In the embodiment of the present invention, since each current accident scene photo will be identified, a first accident type prediction result and a first responsibility allocation result will be output for each current accident scene photo. In addition, since there may be a situation where one picture corresponds to two accident types when training the accident type determination model, the outputted first accident type prediction results may include different accident types. At this time, in order to make the obtained results more accurate, the actual proportion of each accident type in the current accident scene photos will be counted, wherein the actual proportion is calculated based on the number of current accident scene photos corresponding to each of the accident types and the number of current accident scene photos uploaded by the user.

[0076] For example, a user uploads 5 pictures, among which the output first accident type prediction results are rear-end collision of the rear vehicle, reversing of the front vehicle, slipping of the front vehicle, rear-end collision of the rear vehicle and slipping of the front vehicle. At this time, there are 3 types of accidents included, the actual proportion of rear-end collision of the rear vehicle is 40%, the actual proportion of reversing of the front vehicle is 20%, and the actual proportion of slipping of the front vehicle is 40%.

[0077] Step 104: Filter out the largest actual proportion from the actual proportions and compare it with the first preset value.

[0078] The first preset value set in the present invention is set according to actual needs, and the specific data is not limited, and can be 60%, 70% or 80%. After obtaining the actual proportion of each accident type, it is necessary to screen out the largest one. The first accident type prediction result corresponding to the largest actual proportion is closer to the correct result of this identification. However, it is not certain, so it is necessary to compare with the first preset value.

[0079] Step 105: If the maximum actual proportion is greater than the first preset value, the correct second accident type prediction result and the second responsibility allocation result are determined from the first accident type prediction result and the first responsibility allocation result according to the maximum actual proportion.

[0080] In an embodiment of the present invention, if the maximum actual proportion is greater than the first preset value, it is considered that the first accident type prediction result and the first responsibility allocation result corresponding to the maximum actual proportion are representative enough to ignore other identification results, so they are determined as the correct second accident type prediction result and the second responsibility allocation result at this time.

[0081] Step 106: If the maximum actual proportion is less than the first preset value, a first supplementary prompt is sent.

[0082] In the embodiment of the present invention, because there are the same scene photos, but they correspond to completely different accident types and liability determinations, for example, when the front and rear vehicles are close together in the same lane, it may be that the rear vehicle is rear-ending, or the front vehicle is reversing or slipping, so when the current accident scene photo is identified by the pre-trained accident type determination model, multiple prediction results may be output. In reality, one accident corresponds to one accident type, so this output means that the current output result is controversial and needs to be corrected.

[0083] Therefore, in the embodiment of the present invention, by comparing the maximum actual proportion with the first preset value, when the maximum actual proportion is less than the first preset value, it proves that the interference term at this time is too large, and the result identified by the accident type determination model is not representative. At this time, it is assumed that there is doubt about the prediction result of the first accident type or that the prediction result of the first accident type is inconsistent with the facts. At this time, the first supplementary prompt is sent to prompt the user to enter key information to correct the result.

[0084] Step 107 , receiving key information input by the user in response to the first supplementary prompt, and generating a third accident type prediction result and a third responsibility allocation result regarding the accident type of the current accident.

[0085] In the embodiment of the present invention, the supplementary key information may include: the speed of the vehicle at the time of the accident, the speed of other vehicles at the time of the accident, the state of the signal light at the time of the accident, the weather state at the time of the accident, the driving behavior / driving state (drunk driving, etc.) at the time of the accident, the road use state at the time of the accident, etc. After receiving the input of this supplementary information, the cloud server will correspond the input key information to the first key information feature, such as: lane line position, lane line type, road markings, etc., and recombine the key information input by the user with the key feature information extracted from the accident photo, for example: the speed of the vehicle at the time of the accident input by the user is combined with the lane line position in the accident scene photo, the relative position of the vehicle and other feature information, so as to update the accident type prediction, and at this time, the first accident type prediction result and the first responsibility allocation result with the largest proportion of accident types are taken as the third accident type prediction result and responsibility allocation result of the accident type of the current accident. The cloud server will feedback the updated prediction result to the user, and the user can confirm or raise questions again to further adjust the prediction result so that the final result is more reasonable and accurate.

[0086] In addition, the third-party liability allocation results can help users (car owners) more clearly understand the cause and responsibility of the accident, and when it is determined that the other party is responsible, it can help users (car owners) take a more proactive position in the subsequent negotiation process.

[0087] The data processing method provided by the embodiment of the present invention receives a number of current accident scene photos uploaded by a user, and identifies first key feature information in the number of current accident scene photos, so as to facilitate the subsequent identification of the accident type through this key feature information, input the first key feature information into a pre-trained accident type determination model, and output a first accident type prediction result and a first responsibility allocation result about the current accident based on each of the current accident scene photos; obtains the accident type types contained in the first accident type prediction results and the actual proportion of each of the accident type types in the number of current accident scene photos; selects the largest actual proportion from the actual proportions and compares it with a first preset value; if the largest actual proportion is greater than the first preset value, determines the correct second accident type prediction result and the second responsibility allocation result from the first accident type prediction result and the first responsibility allocation result through the largest actual proportion, The largest actual proportion represents the most likely recognition result. If the probability of this result is greater than the first preset value, the representativeness of this result is further determined. If the largest actual proportion is less than the first preset value, a first supplementary prompt is sent, and key information input by the user for the first supplementary prompt is received. A third accident type prediction result and a third responsibility allocation result regarding the accident type of the current accident are generated. Through the user's supplementary input, the output first accident type prediction result is corrected, so that the result obtained is closer to the actual situation of the accident and the responsibility determination is more accurate. The embodiment of the present invention recognizes the accident type to which the accident scene photo belongs and determines the responsibility allocation result through a pre-trained accident type determination model, thereby realizing rapid recording and analysis of traffic accidents, improving the efficiency of accident handling, and avoiding the occurrence of greater safety hazards. By supplementing key information, the rationality of the conclusion is guaranteed and the uncertainty of subjective judgment is reduced.

[0088] Reference Figure 2 , shows a flowchart of another data processing method provided by an embodiment of the present invention, the method may include:

[0089] Step 201: receiving a number of current accident scene photos uploaded by a user, and identifying first key feature information in the number of current accident scene photos.

[0090] Step 202: Input the first key feature information into a pre-trained accident type determination model, and output a first accident type prediction result and a first responsibility allocation result for the current accident based on each current accident scene photo.

[0091] Step 203: Obtain the accident types included in the first accident type prediction result and the actual proportion of each accident type in a number of current accident scene photos.

[0092] Step 204: Filter out the largest actual proportion from the actual proportions and compare it with the first preset value.

[0093] Step 205: If the maximum actual proportion is greater than the first preset value, the correct second accident type prediction result and the second responsibility allocation result are determined from the first accident type prediction result and the first responsibility allocation result according to the maximum actual proportion.

[0094] Step 206: If the maximum actual proportion is less than the first preset value, a first supplementary prompt is sent.

[0095] Step 207: receiving key information input by the user in response to the first supplementary prompt, and generating a third accident type prediction result and a third responsibility allocation result regarding the accident type of the current accident.

[0096] The above steps 201-207 refer to the discussion of the above steps 101-107, and the present invention will not elaborate on them here.

[0097] Step 208, collecting original images of different parts and components of different models of vehicles.

[0098] In the embodiment of the present invention, the damaged parts of the vehicle involved in the accident are also priced to determine the cost of the current vehicle repair. In order to identify the damaged parts of the vehicle, original pictures of parts of different parts of different models of vehicles are collected in advance. These original pictures can be pictures of the vehicle just leaving the factory. After collecting the original pictures, each picture is labeled with the corresponding vehicle model.

[0099] Step 209, compare the original image with the current accident scene photo to determine the second key feature information about the current vehicle in the current accident scene photo, the second key feature information including the current vehicle model, the damaged part of the current vehicle and the damage type of the current damaged part.

[0100] In the embodiment of the present invention, the original image is compared with the current accident scene photo, and a feature matching algorithm, such as SIFT, SURF or a deep learning method, is used to find the most similar original image. Based on the comparison result, the model of the current vehicle is determined, the damaged parts in the accident scene photo are identified, and the damage type of the damaged parts is analyzed, such as scratches, dents, fractures, etc. The determined information is verified to ensure accuracy.

[0101] Step 210, searching a preset vehicle maintenance database through the second key feature information to determine a first maintenance cost for each damaged part of the current vehicle.

[0102] In the embodiment of the present invention, cooperation is carried out with a professional vehicle maintenance database to obtain data on the repair costs of parts for different vehicle models and different damaged parts. In order to ensure that the information in the database is up to date and covers various possible repair situations, the maintenance database will be updated regularly to reflect new models and changes in maintenance costs on the market. The model and system performance will also be continuously optimized based on user feedback and actual application effects.

[0103] When new accident scene photos are received, the damaged parts of the current vehicle and the damage type of the current damaged parts are identified. According to the identification results, the maintenance database is queried to obtain the corresponding parts repair costs. Among them, different repair costs correspond to different types of damage to the same damaged part of the same vehicle model, and different repair costs correspond to the same damaged part of different vehicle models and the same type of damage.

[0104] Step 211, determining the target number of damaged parts of the current vehicle through the second key feature information.

[0105] In the above process of the embodiment of the present invention, what is identified is the first repair cost of each damaged part of the current vehicle, but all damaged parts of the vehicle will generate corresponding first repair costs, so it is necessary to determine the number of damaged parts of the current vehicle to facilitate the subsequent acquisition of the corresponding number of first repair costs, and then add up these first repair costs to obtain the final repair cost.

[0106] Step 212: The target number of first maintenance costs is accumulated to generate a second maintenance cost for the current vehicle.

[0107] The embodiment of the present invention accumulates the first maintenance costs of the target number to generate the second maintenance cost of the current vehicle. For example, if it is detected that there are three damaged parts, then it is necessary to obtain the first maintenance costs of the three parts, and then accumulate them to obtain the cost required for the vehicle to be repaired this time.

[0108] It should be noted that although image recognition and database query can provide a preliminary estimate of repair costs, some unforeseen problems may be encountered during the actual repair process, which may affect the final repair cost. Therefore, the estimate results provided to users should be clearly marked as reference in nature, and users are advised to take the following measures: When providing the estimate results, clearly inform users that the results are based on the current visible damage and information in the database, and may not include all potential repair needs. Emphasize that the estimate results are for reference only and the actual repair costs may vary; recommend users to send the vehicle to a professional repair organization for a detailed inspection to identify all damaged parts and potential problems, and remind users that only after inspection by professional technicians can an accurate repair quotation be obtained; if possible, provide a list of reliable repair organizations near the user to facilitate user selection, and establish cooperative relationships with these repair organizations to ensure that users can receive professional services; provide follow-up tracking services, ask users whether they have conducted professional inspections and understand the actual repair costs, and continuously optimize database information based on user feedback; explain factors that may affect repair costs, such as the availability of parts, difficulty of repair, market price fluctuations, etc.

[0109] These measures ensure that users have reasonable expectations of the estimated results and understand the importance of professional inspections before making maintenance decisions. At the same time, this also reflects the professionalism of the service provider and its responsible attitude towards users.

[0110] Reference Figure 3 , shows a flowchart of another data processing method provided by an embodiment of the present invention, the method may include:

[0111] Step 301, receiving a current accident scene photo uploaded by a user, and identifying the first key feature information in the current accident scene photo.

[0112] Step 302: Input the first key feature information into a pre-trained accident type determination model, and output a first accident type prediction result and a first responsibility allocation result for the current accident based on each current accident scene photo.

[0113] Step 303: Obtain the accident types included in the first accident type prediction result and the actual proportion of each accident type in a number of current accident scene photos.

[0114] Step 304: Filter out the largest actual proportion from the actual proportions and compare it with the first preset value.

[0115] Step 305: If the maximum actual proportion is greater than the first preset value, the correct second accident type prediction result and the second responsibility allocation result are determined from the first accident type prediction result and the first responsibility allocation result according to the maximum actual proportion.

[0116] Step 306: If the maximum actual proportion is less than the first preset value, a first supplementary prompt is sent.

[0117] Step 307: receiving key information input by the user in response to the first supplementary prompt, and generating a third accident type prediction result and a third responsibility allocation result regarding the accident type of the current accident.

[0118] The above steps 301-307 refer to the discussion of the above steps 101-107, and the present invention will not elaborate on them here.

[0119] Step 308, obtaining the user's insurance information for the current vehicle.

[0120] In the embodiment of the present invention, the insurance information of the vehicle is also obtained to determine the insurance claim amount that the owner can obtain in the current accident and the increased amount when the next insurance is purchased, and then the most appropriate solution is determined based on these.

[0121] Step 309, determining the target conditions for the insurance policy to take effect through the insurance information.

[0122] Since there are many types of vehicle insurance, the insurance information obtained in the embodiment of the present invention may be more than one type, which may include compulsory traffic insurance, commercial third party liability insurance, vehicle loss insurance, driver and passenger seat liability insurance, glass breakage insurance, no deductible special insurance, body scratch insurance, designated repair shop insurance, new equipment loss insurance and mental damage consolation money liability insurance, etc. Each type of vehicle insurance corresponds to different insurance information, and the effective target conditions are also different.

[0123] Step 310, determining whether the third accident type prediction result and the third responsibility allocation result meet the target conditions.

[0124] In the embodiment of the present invention, whether the occurrence condition is met is determined by obtaining the third accident type prediction result and the third responsibility allocation result.

[0125] Step 311, if satisfied, determine the target policy information corresponding to the target condition.

[0126] If the accident conditions are met, the corresponding target insurance policy information needs to be determined based on the target conditions met. For example, the current vehicle contains 7 types of insurance information and 7 corresponding target conditions, but the vehicle only meets 3 of the conditions. In this case, the corresponding target insurance policy information of the 3 target conditions needs to be obtained.

[0127] Step 312, determining the target compensation amount obtained through the target policy information.

[0128] In the embodiment of the present invention, the target compensation amount is determined by the target insurance policy information. It should be noted that this target compensation amount is an estimated result, and the actual compensation result needs to be determined after the subsequent insurance personnel arrive at the scene.

[0129] It should be noted that, in addition to considering the insurance compensation amount, the present invention also considers the additional amount of insurance purchased next time after the insurance claim is settled this time, and makes a compensation suggestion after comprehensive consideration. The specific steps include:

[0130] Retrieve in a pre-set vehicle maintenance database through the third accident type prediction result, the third responsibility allocation result and the target maintenance cost of the current vehicle;

[0131] Determine the additional premium amount for reinsurance through the search results. The additional premium amount is the additional cost for reinsurance over the current insurance for the same type of insurance;

[0132] Compare the target claim amount with the premium increase amount;

[0133] Generate reasonable compensation recommendations by comparing the results.

[0134] For example, if the car owner in the current accident obtains a claim amount of 1,000 yuan through insurance claim, but the additional amount required for the next insurance is 1,500 yuan, because 1,500>1,000, it can be determined that it is not cost-effective to claim through insurance. At this time, you can negotiate with the other car owner and make a private monetary payment instead of insurance claim and repair at your own expense. If the additional amount required for the next insurance is 700 yuan, because 700<1,000, it can be determined that it is cost-effective to claim through insurance. At this time, you can notify the insurance company and wait for the insurance personnel to arrive to make the insurance claim, helping users make wise decisions.

[0135] It should be noted that when predicting the increase in premium for reinsurance, in order to improve the accuracy of the prediction, it is necessary to establish a data table containing historical accident data (including accident type), responsibility division, maintenance costs, premium changes and other information in the vehicle maintenance database, and then search the data table through the information of the current accident to determine the premium change. It is also possible to train the model to predict the premium change after the accident through machine learning algorithms, such as linear regression, decision trees, etc., which are not specifically limited in the present invention. Among them, the predicted increase in premium for reinsurance will be personalized according to different insurance companies and insurance terms to provide users with more accurate premium change estimates.

[0136] This application also identifies the damaged parts of the vehicle and combines it with a pre-set vehicle maintenance database to determine the vehicle's repair costs and obtain the vehicle's insurance status. Based on the analysis of the insurance status and repair costs, the car owner can clearly and quickly obtain accident handling suggestions and insurance claim details, which not only helps the car owner make more information-supported decisions, but also reduces waiting time and alleviates road congestion. Coupled with direct evidence support and smart recommendations, it simplifies the claims process and improves the success rate of claims.

[0137] Reference Figure 4 , which shows a schematic diagram of the structure of a data processing device provided by an embodiment of the present invention, such as Figure 4 As shown, the device may include:

[0138] The first recognition module 401 is used to receive a number of current accident scene photos uploaded by a user, and recognize the first key feature information in the number of current accident scene photos.

[0139] The first output module 402 is used to input the first key feature information into a pre-trained accident type determination model, and output a first accident type prediction result and a first responsibility allocation result about the current accident based on each current accident scene photo.

[0140] The first acquisition module 403 is used to obtain the accident type categories included in the first accident type prediction result and the actual proportion of each accident type category in a number of current accident scene photos.

[0141] The first screening module 404 is used to screen out the largest actual proportion from the actual proportions and compare it with the first preset value.

[0142] The first determination module 405 is used to determine the correct second accident type prediction result and the second responsibility allocation result from the first accident type prediction result and the first responsibility allocation result according to the maximum actual proportion if the maximum actual proportion is greater than the first preset value.

[0143] The first sending module 406 is configured to send a first supplementary prompt if the maximum actual proportion is less than a first preset proportion.

[0144] The first generating module 407 is used to receive key information input by the user in response to the first supplementary prompt, and generate a third accident type prediction result and a third responsibility allocation result regarding the accident type of the current accident.

[0145] Optionally, the data processing device specifically includes:

[0146] The first acquisition module is used to acquire at least one historical accident scene photo.

[0147] The second acquisition module is used to annotate each historical accident scene photo with the corresponding accident type and responsibility allocation result and obtain the key feature information of each historical accident scene photo. The key feature information includes road information and vehicle information. The accident types include rear-end collisions, overtaking and lane changing accidents, reversing and sliding accidents, intersections and roundabouts accidents, and dedicated lane accidents.

[0148] A relationship building module is used to establish an association relationship between the accident type, responsibility allocation results and key feature information of each historical accident scene photo.

[0149] A training module is used to train an accident type determination model through at least one association relationship, wherein key feature information in each association relationship is used as input information, and the accident type and responsibility allocation result are used as corresponding output information.

[0150] The second sending module is used to send a second supplementary prompt if it is detected that the amount of the identified first key feature information is less than a second preset value.

[0151] The updating module is used to update the first key feature information in response to the user's supplementary input for the second supplementary prompt.

[0152] The second acquisition module is used to collect original images of different parts and components of different models of vehicles.

[0153] The second determination module is used to compare the original image with the current accident scene photo to determine the second key feature information about the current vehicle in the current accident scene photo, the second key feature information including the current vehicle model, the damaged part of the current vehicle and the damage type of the current damaged part.

[0154] The third determination module is used to search in a preset vehicle maintenance database through the second key feature information to determine the first repair cost of each damaged part of the current vehicle, wherein different repair costs correspond to different damage types of the same damaged part of the same vehicle model, and different repair costs correspond to the same damaged part of different vehicle models and the same damage type.

[0155] The fourth determination module is used to determine the target number of damaged parts of the current vehicle through the second key feature information.

[0156] The second generating module is used to accumulate the first maintenance costs of the target quantity and generate the second maintenance cost of the current vehicle.

[0157] The third acquisition module is used to obtain the user's insurance information for the current vehicle.

[0158] The fifth determination module is used to determine the target conditions for the effectiveness of the insurance policy through the insurance information.

[0159] The judgment module is used to judge whether the third accident type prediction result and the third responsibility allocation result meet the target conditions.

[0160] The sixth determination module is used to determine the target policy information corresponding to the target condition if it is met.

[0161] The seventh determination module is used to determine the target compensation amount obtained through the target policy information.

[0162] The retrieval module is used to search a preset vehicle maintenance database according to the third accident type prediction result, the third responsibility allocation result and the target maintenance cost of the current vehicle.

[0163] The eighth determination module is used to determine the additional amount of premium for reinsurance through the search results. The additional amount of premium is the additional fee paid for reinsurance than the current insurance for the same type of insurance.

[0164] The comparison module is used to compare the target compensation amount with the premium increase amount.

[0165] The third generation module is used to generate reasonable compensation suggestions based on the comparison results.

[0166] The data processing method provided by the embodiment of the present invention receives a number of current accident scene photos uploaded by a user, and identifies the first key feature information in the number of current accident scene photos, so as to facilitate the subsequent identification of the accident type through this key feature information, input the first key feature information into a pre-trained accident type determination model, and output a first accident type prediction result and a first responsibility allocation result about the current accident based on each of the current accident scene photos; obtains the accident type types contained in the first accident type prediction result and the actual proportion of each of the accident type types in the number of current accident scene photos; selects the largest actual proportion from the actual proportions and compares it with a first preset value; if the largest actual proportion is greater than the first preset value, determines the correct second accident type prediction result and the second responsibility allocation result from the first accident type prediction result and the first responsibility allocation result through the largest actual proportion, and takes the largest actual proportion to represent the most likely recognition result. If the probability of this result is greater than the first preset value, the representativeness of this result is further determined. If the largest actual proportion is less than the first preset value, a first supplementary prompt is sent, and the user's response to the The key information input by the first supplementary prompt is used to generate a third accident type prediction result and a third responsibility allocation result about the accident type of the current accident. Through the user's supplementary input, the output first accident type prediction result is corrected, so that the obtained result is closer to the actual situation of the accident and the responsibility identification is more accurate. The embodiment of the present invention recognizes the accident type to which the accident scene photo belongs and determines the responsibility allocation result through a pre-trained accident type determination model, thereby realizing rapid recording and analysis of traffic accidents, improving the handling efficiency of accidents, and avoiding the occurrence of greater safety hazards. By supplementing key information, the rationality of the conclusion is guaranteed and the uncertainty of subjective judgment is reduced. In addition, the present application also determines the vehicle's maintenance costs and obtains the vehicle's insurance status by identifying the components of the damaged parts of the vehicle in combination with a pre-set vehicle maintenance database. Based on the analysis of the insurance status and maintenance costs, the car owner can clearly and quickly obtain accident handling suggestions and insurance claim details, which not only helps the car owner make more information-supported decisions, but also reduces waiting time and alleviates road congestion. Coupled with direct evidence support and intelligent recommendations, the claim process is simplified and the claim success rate is improved.

[0167] The embodiment of the present invention also provides a communication device, such as Figure 5 As shown, it includes a processor 501, a communication interface 502, a memory 503 and a communication bus 504, wherein the processor 501, the communication interface 502, and the memory 503 communicate with each other through the communication bus 504.

[0168] Memory 503, used for storing computer programs;

[0169] The processor 501 is used to execute the program stored in the memory 503, and implements the following steps:

[0170] Receiving a number of current accident scene photos uploaded by a user, and identifying first key feature information in the number of current accident scene photos;

[0171] Inputting the first key feature information into a pre-trained accident type determination model, and outputting a first accident type prediction result and a first responsibility allocation result for the current accident based on each of the current accident scene photos;

[0172] Obtaining the types of accidents included in the first accident type prediction result and the actual proportion of each of the accident types in a number of current accident scene photos;

[0173] Filter out the largest actual proportion from the actual proportions and compare it with the first preset value;

[0174] If the maximum actual proportion is greater than the first preset value, determining the correct second accident type prediction result and the second responsibility allocation result from the first accident type prediction result and the first responsibility allocation result according to the maximum actual proportion;

[0175] If the maximum actual proportion is less than the first preset value, sending a first supplementary prompt;

[0176] Receive key information input by the user in response to the first supplementary prompt, and generate a third accident type prediction result and a third responsibility allocation result regarding the current accident.

[0177] The communication bus mentioned in the above terminal can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0178] The communication interface is used for communication between the above terminal and other devices.

[0179] The memory may include a random access memory (RAM) or a non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0180] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0181] The present invention also provides a readable storage medium, when the instructions in the storage medium are executed by a processor of an electronic device, the electronic device can execute the data processing method of the aforementioned embodiment.

[0182] Based on the same inventive concept, an embodiment of the present invention further provides a vehicle, which may specifically include: the above-mentioned data processing device.

[0183] As for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0184] The algorithm and display provided herein are not inherently related to any particular computer, virtual device or other equipment. According to the above description, it is obvious that the structure required for constructing this type of device is. In addition, the present invention is not directed to any specific programming language. It should be understood that various programming languages ​​can be utilized to realize the content of the present invention described herein, and the description of the above specific language is to disclose the best mode of the present invention.

[0185] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. In some instances, well-known methods, structures and techniques are not shown in detail so as not to obscure the understanding of this description.

[0186] Similarly, it should be understood that in order to streamline the present invention and aid in understanding one or more of the various inventive aspects, in the above description of exemplary embodiments of the present invention, various features of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. However, this disclosed method should not be interpreted as reflecting the following intention: that the claimed invention requires more features than those expressly recited in each claim. Rather, as reflected in the claims below, inventive aspects lie in less than all of the features of the individual embodiments disclosed above. Therefore, the claims that follow the detailed description are hereby expressly incorporated into the detailed description, with each claim itself serving as a separate embodiment of the present invention.

[0187] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and in addition they may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive, all features disclosed in this specification (including the accompanying claims, abstracts and drawings) and all processes or units of any method or device disclosed in this manner may be combined in any combination. Unless otherwise expressly stated, each feature disclosed in this specification (including the accompanying claims, abstracts and drawings) may be replaced by an alternative feature providing the same, equivalent or similar purpose.

[0188] The various component embodiments of the present invention may be implemented in hardware, or in software modules running on one or more processors, or in a combination thereof. It should be understood by those skilled in the art that a microprocessor or digital signal processor (DSP) may be used in practice to implement some or all of the functions of some or all of the components in the sorting device according to the present invention. The present invention may also be implemented as a device or apparatus program for executing part or all of the methods described herein. Such a program for implementing the present invention may be stored on a computer-readable medium, or may be in the form of one or more signals. Such a signal may be downloaded from an Internet website, or provided on a carrier signal, or provided in any other form.

[0189] It should be noted that the above embodiments illustrate the present invention rather than limit it, and that those skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets shall not be construed as a limitation on the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention may be implemented by means of hardware comprising a number of different elements and by means of a suitably programmed computer. In a unit claim enumerating a number of devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc., does not indicate any order. These words may be interpreted as names.

[0190] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices, apparatuses and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0191] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

[0192] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0193] It should be noted that the various data-related processes in the embodiments of the present application are all carried out in compliance with the corresponding data protection laws and policies of the country where the device is located, and with the authorization given by the owner of the corresponding device.

Claims

1. A data processing method, characterized in that: The method comprises: Receiving a number of current accident scene photos uploaded by a user, and identifying first key feature information in the number of current accident scene photos; Inputting the first key feature information into a pre-trained accident type determination model, and outputting a first accident type prediction result and a first responsibility allocation result for the current accident based on each of the current accident scene photos; Obtaining the types of accidents included in the first accident type prediction result and the actual proportion of each of the accident types in a number of current accident scene photos; Filter out the largest actual proportion from the actual proportions and compare it with the first preset value; If the maximum actual proportion is greater than the first preset value, determining the correct second accident type prediction result and the second responsibility allocation result from the first accident type prediction result and the first responsibility allocation result according to the maximum actual proportion; If the maximum actual proportion is less than the first preset value, sending a first supplementary prompt; Receive key information input by the user in response to the first supplementary prompt, and generate a third accident type prediction result and a third responsibility allocation result regarding the current accident.

2. The data processing method according to claim 1, characterized in that: Before receiving a plurality of current accident scene photos uploaded by a user and identifying the first key feature information in the plurality of current accident scene photos, the method further includes: Collect at least one historical accident scene photo; Marking each of the historical accident scene photos with the corresponding accident type and responsibility allocation result and obtaining key feature information of each of the historical accident scene photos, the key feature information including road information and vehicle information, and the accident types including rear-end collisions, overtaking and lane change accidents, reversing and rolling accidents, intersection and roundabout accidents, and dedicated lane accidents; Establishing a correlation relationship between the accident type, responsibility allocation results and key feature information of each of the historical accident scene photos; The accident type determination model is trained through at least one of the association relationships, wherein key feature information in each association relationship is used as input information, and the accident type and responsibility allocation result are used as corresponding output information.

3. The data processing method according to claim 1, characterized in that: Before inputting the first key feature information into a pre-trained accident type determination model and outputting a first accident type prediction result and a first responsibility allocation result for the current accident based on each current accident scene photo, the method further includes: If it is detected that the amount of the first key feature information identified is less than the second preset value, sending a second supplementary prompt; In response to the user's supplementary input for the second supplementary prompt, the first key feature information is updated.

4. The data processing method according to claim 1, characterized in that: After receiving the key information input by the user in response to the first supplementary prompt and generating a third accident type prediction result and a third responsibility allocation result about the current accident, the method further includes: Collect original pictures of different parts of different models of vehicles; Comparing the original image with the current accident scene photo to determine second key feature information about the current vehicle in the current accident scene photo, the second key feature information including the current vehicle model, the damaged part of the current vehicle, and the damage type of the current damaged part; The first repair cost of each damaged part of the current vehicle is determined by searching the preset vehicle repair database through the second key feature information, wherein different repair costs correspond to different types of damage to the same damaged part of the same vehicle model, and different repair costs correspond to the same damaged part of different vehicle models and the same type of damage to the same vehicle model; Determine the target number of damaged parts of the current vehicle through the second key feature information; The target number of the first maintenance costs is accumulated to generate the second maintenance cost of the current vehicle.

5. The data processing method according to claim 1, characterized in that: After receiving the key information input by the user in response to the first supplementary prompt and generating a third accident type prediction result and a third responsibility allocation result regarding the accident type of the current accident, the method further includes: Get the user's insurance information for the current vehicle; Determine the target conditions for the effectiveness of the insurance policy through the insurance information; Determining whether the third accident type prediction result and the third responsibility allocation result meet the target condition; If satisfied, then determine the target policy information corresponding to the target condition; The target compensation amount is determined by the target policy information.

6. The data processing method according to claim 5, characterized in that: After the target compensation amount is determined through the target insurance policy information, the method further includes: Retrieving the third accident type prediction result, the third responsibility allocation result and the target repair cost for the current vehicle from a preset vehicle repair database; Determine the additional amount of premium for reinsurance through the search results, wherein the additional amount of premium is the additional cost of reinsurance over the current insurance for the same type of insurance; comparing the target compensation amount and the premium increase amount; Generate reasonable compensation recommendations by comparing the results.

7. A data processing device, characterized in that: include: A first recognition module is used to receive a number of current accident scene photos uploaded by a user, and recognize first key feature information in the number of current accident scene photos; A first output module, configured to input the first key feature information into a pre-trained accident type determination model, and output a first accident type prediction result and a first responsibility allocation result for the current accident based on each of the current accident scene photos; A first acquisition module, used to acquire the types of accidents included in the first accident type prediction result and the actual proportion of each of the accident types in a number of current accident scene photos; A first screening module, used to screen out the largest actual proportion from the actual proportions and compare it with a first preset value; A first determination module, configured to determine a correct second accident type prediction result and a second responsibility allocation result from the first accident type prediction result and the first responsibility allocation result by using the maximum actual proportion if the maximum actual proportion is greater than a first preset value; A first sending module, configured to send a first supplementary prompt if the maximum actual proportion is less than a first preset value; The first generating module is used to receive key information input by the user in response to the first supplementary prompt, and generate a third accident type prediction result and a third responsibility allocation result regarding the accident type of the current accident.

8. A communication device, characterized in that: include: A transceiver, a memory, a processor, and a program stored on the memory and executable on the processor; The processor is used to read the program in the memory to implement the steps in the data processing method as described in any one of claims 1-6.

9. A readable storage medium for storing a program, characterized in that: When the stored program is executed by a processor, the steps of the data processing method according to any one of claims 1 to 6 are implemented.

10. A vehicle, characterized in that: include: The data processing device as claimed in claim 7.