Vehicle accident handling method, device, equipment and medium

Abstract

The invention relates to a vehicle accident handling method, device and equipment and a medium, and the method comprises the steps: obtaining vehicle driving data; performing accident prediction on the vehicle driving data based on a pre-trained identification model to obtain an accident prediction result; the accident prediction result represents whether a potential accident exists or not; acquiring sensor data under the condition that the accident prediction result is that a potential accident exists, and matching with a preset accident condition according to the sensor data to determine an accident identification result; and when the accident identification result is that the vehicle accident occurs, generating an accident report of the vehicle accident. According to the technical scheme, the vehicle accident can be accurately recognized, the accident report can be generated, and the vehicle accident handling accuracy and the accident handling efficiency are improved.

Description

Technical Field

[0001] The present disclosure relates to the field of vehicle technology, and in particular to a vehicle accident handling method, device, equipment and medium. Background Art

[0002] With the increase in the number of vehicles and the increasing complexity of modern road environments, the frequency and complexity of traffic accidents are also increasing. Identifying vehicle accidents is of great significance for accident handling, which usually involves preserving the accident scene, collecting evidence, determining liability and damages, and insurance claims. It is not only time-consuming and labor-intensive for users, but may also affect the smooth flow of roads. Therefore, there is a need to improve the accuracy and efficiency of accident handling.

[0003] At present, when dealing with vehicle accidents, there are solutions for automatically identifying vehicle accidents for accident handling. However, this solution still has the problem of missed vehicle accidents, which affects the accuracy of accident handling, and the efficiency of accident handling still needs to be improved. How to handle vehicle accidents accurately and efficiently is a technical problem that needs to be solved urgently. Summary of the invention

[0004] In order to solve the above technical problems, the present disclosure provides a vehicle accident handling method, device, equipment and medium.

[0005] In a first aspect, an embodiment of the present disclosure provides a method for handling a vehicle accident, comprising:

[0006] Obtain vehicle driving data;

[0007] Performing accident prediction on the vehicle driving data based on a pre-trained recognition model to obtain an accident prediction result; the accident prediction result indicates whether there is a potential accident;

[0008] When the accident prediction result indicates that there is a potential accident, obtaining sensor data, and matching the sensor data with a preset accident condition to determine an accident identification result;

[0009] When the accident identification result is that a vehicle accident has occurred, an accident report of the vehicle accident is generated.

[0010] In a second aspect, an embodiment of the present disclosure provides a vehicle accident handling device, comprising:

[0011] An acquisition module, used for acquiring vehicle driving data;

[0012] A prediction module, used to perform accident prediction on the vehicle driving data based on a pre-trained recognition model to obtain an accident prediction result; the accident prediction result indicates whether there is a potential accident;

[0013] A determination module, used for obtaining sensor data when the accident prediction result indicates that a potential accident exists, and matching the sensor data with a preset accident condition to determine an accident identification result;

[0014] A generating module is used to generate an accident report of the vehicle accident when the accident identification result is that a vehicle accident has occurred.

[0015] In a third aspect, an embodiment of the present disclosure provides an electronic device, comprising: a processor; a memory for storing executable instructions of the processor; the processor is used to read the executable instructions from the memory and execute the instructions to implement the vehicle accident handling method described in the first aspect above.

[0016] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the vehicle accident handling method described in the first aspect is implemented.

[0017] The technical solution provided by the disclosed embodiment has the following advantages over the prior art: vehicle driving data is predicted through a pre-trained recognition model to obtain an accident prediction result, and then, when the accident prediction result is the existence of a potential accident, the sensor data is matched with the preset accident conditions to determine the accident recognition result, and further, when the accident recognition result is the occurrence of a vehicle accident, an accident report of the vehicle accident is generated. Thus, potential accidents are identified based on vehicle driving data, and then the potential accidents are confirmed based on the sensor data, which can accurately identify vehicle accidents and avoid missed vehicle accidents, and then accurately generate accident reports based on the vehicle accidents for accident handling, thereby improving the accuracy and efficiency of accident handling. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0019] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0020] Figure 1 A schematic diagram of a flow chart of a vehicle accident handling method provided by an embodiment of the present disclosure;

[0021] Figure 2A schematic flow chart of another vehicle accident handling method provided by an embodiment of the present disclosure;

[0022] Figure 3 A schematic diagram of a process for identifying fraudulent behavior in vehicle accidents provided by an embodiment of the present disclosure;

[0023] Figure 4 A schematic flow chart of another vehicle accident handling method provided by an embodiment of the present disclosure;

[0024] Figure 5 A schematic structural diagram of a vehicle accident handling device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0025] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features in the embodiments can be combined with each other without conflict.

[0026] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.

[0027] Figure 1 This is a flow chart of a vehicle accident handling method provided in an embodiment of the present disclosure. The method provided in the embodiment of the present disclosure can be executed by a vehicle accident handling device, which can be implemented using software and / or hardware and can be integrated on any electronic device with computing capabilities.

[0028] like Figure 1 As shown, the vehicle accident handling method provided by the embodiment of the present disclosure may include:

[0029] Step 101, obtaining vehicle driving data.

[0030] In this embodiment, the vehicle driving data is acquired during the vehicle driving process, wherein the vehicle driving data may be collected periodically or when certain conditions are met.

[0031] In one embodiment of the present disclosure, obtaining vehicle driving data includes: obtaining vehicle driving data at a first acquisition frequency according to a preset cycle; and / or, in response to detecting an abnormal vehicle driving event, obtaining vehicle driving data at a second acquisition frequency in a preset time period. Wherein, the second acquisition frequency is greater than the first acquisition frequency. Since a vehicle accident is often accompanied by an abnormal vehicle driving event, when an abnormal vehicle driving event is detected, the vehicle driving data is obtained at a second acquisition frequency that is higher than the first acquisition frequency, so that the vehicle driving data before the accident can be obtained more comprehensively and accurately. When no abnormal vehicle driving event is detected, the second acquisition frequency is not used for acquisition, and processing resource consumption is taken into account. In this embodiment, abnormal vehicle driving events include but are not limited to vibration events, emergency stop events, airbag triggering events, etc. For example, taking an emergency stop event as an example, when it is detected that the rate of change of the vehicle speed reduction within a preset time period is greater than a threshold, it is determined that an emergency stop event has occurred. Wherein, the length of the preset time period can be set according to the actual scenario, and is not limited here.

[0032] As an example, data collection is set to be performed periodically, and vehicle driving data is collected every 100 milliseconds to achieve real-time monitoring of the vehicle status.

[0033] As another example, when an abnormal vehicle driving event is detected, data collection is triggered and vehicle driving data is collected every 10 milliseconds. In this example, the data collection is not limited by the normal collection cycle to ensure the integrity and accuracy of the data before and after the accident. The preset time period can be set to ten minutes to provide sufficient context information.

[0034] In this embodiment, the acquisition of vehicle driving data can be realized based on a data acquisition module, which is integrated into the vehicle intelligent driving assistance system and the vehicle networking platform to collect key data during vehicle driving in real time. The data acquisition module can include an analysis algorithm for determining whether the vehicle status is normal. When a potential safety problem is detected, a warning can be issued to the driver or measures can be automatically taken, such as initiating emergency braking.

[0035] The vehicle driving data includes but is not limited to speed, position, braking status, acceleration, and steering angle. The following is an example of the collection process of the above vehicle driving data. For example, the vehicle speed is calculated by the wheel speed sensor or positioning data, and the speed can be collected according to the set period; the vehicle position is collected by the positioning system, and the position can be collected according to the positioning system update frequency or the set period; the vehicle braking status is detected by the sensor on the brake pedal, and the braking status can be collected according to the set period, or when the braking status changes; the acceleration of the vehicle is measured by the accelerometer, and the acceleration can be collected according to the set period, or when abnormal acceleration is detected; the steering angle of the vehicle is measured by the steering sensor, and the steering angle can be collected according to the set period, or when the steering angle changes.

[0036] Optionally, the collected vehicle driving data can be stored in a storage device inside the vehicle, or sent to a cloud server for storage when the vehicle is connected to the Internet of Vehicles platform. Data storage methods include but are not limited to encrypted storage.

[0037] Step 102 , performing accident prediction on the vehicle driving data based on the pre-trained recognition model to obtain an accident prediction result.

[0038] In this embodiment, for the pre-trained recognition model, the input of the model is a time series of vehicle driving data, and the output is the accident prediction result, which indicates whether there is a potential accident. Accident prediction through vehicle driving data can predict potential vehicle accidents. Accident prediction uses vehicle driving data, which occupies less resources than image processing and other methods. Subsequent confirmation processing only needs to be performed for the predicted potential accidents. Further confirmation based on potential accidents can accurately identify vehicle accidents. For example, when the airbag is not detected to be triggered, vehicle accidents such as minor scratches can be identified to avoid missed accidents.

[0039] The recognition model can be implemented using machine learning or deep learning algorithms, including decision trees, support vector machines, random forests, neural networks, etc., to identify abnormal patterns in vehicle driving data, which indicate that an accident may be about to occur. The training process of the recognition model is as follows: obtain training data, the training data includes first sample driving data within a preset time period before and after the accident, and second sample driving data within a preset time period when no accident occurred. Then, the first sample driving data is used as a positive sample and the second sample driving data is used as a negative sample to train the recognition model, so that the input of the recognition model is a time series of vehicle driving data such as speed, position, acceleration, etc., and the output is whether there is a potential accident.

[0040] As an example, the collected vehicle driving data is preprocessed, which includes data cleaning and data standardization, such as removing outliers and filling missing values. By synchronizing data from different sources and types, the timestamps of the data are ensured to be consistent. Then, key features are extracted from the preprocessed data, such as features in the time domain and frequency domain, and sliding window analysis is performed on the data to capture dynamic behaviors before the accident. Accident prediction is performed using a pre-trained recognition model. The accident prediction results include the presence of potential accidents and the absence of potential accidents. When the accident prediction result is that there is a potential accident, warning information can also be generated to prompt the driver to take risk avoidance measures.

[0041] Step 103, when the accident prediction result indicates that there is a potential accident, sensor data is acquired, and the sensor data is matched with preset accident conditions to determine an accident identification result.

[0042] In this embodiment, when the accident prediction result is that there is a potential accident, sensor data collected by the vehicle sensor is further obtained to determine the accident identification result according to the sensor data. The sensor includes a collision signal sensor, an accelerometer, a camera and a radar, and the sensor data includes a vehicle collision signal, acceleration data, image data, and radar data. When the accident prediction result is that there is no potential accident, the steps of obtaining sensor data and matching the sensor data with the accident conditions are not performed.

[0043] Among them, the accident condition is used to further confirm whether an accident has occurred. The accident identification results include whether a vehicle accident has occurred or not. When the sensor data meets the preset accident conditions, the accident identification result is determined to be that a vehicle accident has occurred.

[0044] Optionally, the accident condition includes: when the vehicle collision signal is detected to be greater than or equal to the collision threshold, or when the vehicle initiates a rescue signal, the accident identification result is determined to be a vehicle accident; and / or, the probability of an accident is determined based on the vehicle acceleration change rate within a specified time period, the collision detection result of the environmental perception image, and the collision detection result of the radar data. If the probability of an accident is greater than a preset threshold, the accident identification result is determined to be a vehicle accident. It should be noted that the above-mentioned implementation method of matching sensor data with accident conditions is only one possible implementation method, and it can also be matched in combination with multiple information such as driver reaction and airbag status, and the above-mentioned implementation method can be implemented alone or in multiple combinations, which is not limited here.

[0045] As an example, when the accident prediction result is that there is a potential accident, the vehicle collision signal, acceleration data, image data, radar data and other sensor data can be collected by collision signal sensors, accelerometers, cameras and radars at the same time, and then the sensor data is matched with the preset accident conditions to determine the accident identification result. In this example, when any sensor data meets the accident condition, it is determined that a vehicle accident has occurred, and when all sensor data do not meet the accident condition, it is determined that no vehicle accident has occurred. For example, when the vehicle collision signal is detected to be greater than 0.6, or when the vehicle initiates a rescue signal, or according to the vehicle acceleration change rate within a specified time period, the collision detection result of the environmental perception image, and the collision detection result of the radar data, the probability of the accident is determined. When the probability of the accident is greater than the preset threshold of 0.7, it is determined that a vehicle accident has occurred; if the vehicle collision signal is less than or equal to 0.6, and the vehicle does not initiate a rescue signal, and the probability of the accident is less than or equal to the preset threshold of 0.7, it is determined that no vehicle accident has occurred.

[0046] As an example, when the accident prediction result is that there is a potential accident, the vehicle collision signal can be first collected through the collision signal sensor, and the accident identification result can be determined according to the vehicle collision signal. Then, when the vehicle collision signal does not meet the accident condition, the rescue signal is further detected, and / or, the acceleration data, image data, radar data and other sensor data are collected through sensors such as accelerometers, cameras and radars to determine the probability of the accident. The vehicle collision signal matching method is more efficient than other sensor data such as image data and radar data. Therefore, the accident identification result is first determined by the vehicle collision signal, and when the vehicle collision signal does not meet the accident condition, the accident identification result is further determined by other sensor data. Among them, if the rescue signal meets the accident condition, and / or the probability of the accident meets the accident condition, it is determined that a vehicle accident has occurred. If both the rescue signal and the probability of the accident do not meet the accident condition, it is determined that no vehicle accident has occurred. For example, first compare the vehicle collision signal with the collision threshold, the collision threshold is, for example, 0.6, and when the vehicle collision signal is greater than 0.6, it is determined that a vehicle accident has occurred. If the vehicle collision signal is less than or equal to 0.6, at this time, if it is detected that the vehicle initiates a rescue signal, it is determined that a vehicle accident has occurred. If the vehicle collision signal is less than or equal to 0.6, the probability of an accident is determined based on the vehicle acceleration change rate within a specified time period, the collision detection result of the environmental perception image, and the collision detection result of the radar data. For acceleration data, the vehicle acceleration change rate within a specified time period is obtained, for example, the vehicle acceleration change rate within 3 seconds before and after the vehicle collision signal is generated is obtained. If the vehicle acceleration change rate is greater than a threshold, it is determined that a collision may occur; for image data, the vehicle collects environmental perception images of the vehicle's surrounding environment through a camera, and analyzes whether there are signs of collision through an image processing algorithm. For example, when a deformation of an object in front of the vehicle is detected in the environmental perception image, it is determined that a collision may occur; for radar data, the vehicle detects surrounding obstacles through a radar sensor and determines the change in the distance between the vehicle and the obstacle. When it is detected that the distance between the vehicle and the obstacle is decreasing, and the difference between the rate of distance reduction and the reference rate is greater than a threshold, for example, the difference is greater than 50%, it is determined that a collision may occur. Optionally, the index analysis results of the acceleration data, image data and radar data can be set with different weights, for example, the weight of acceleration data is 0.6, the weight of image data is 0.2, and the weight of radar data is 0.2. The analysis results are combined with the weights for weighted summation to determine the probability of an accident. Further, if the probability of an accident is greater than a preset threshold of 0.7, it is determined that a vehicle accident has occurred.

[0047] Among them, the accident probability is determined through a comprehensive analysis of three indicators: acceleration data, image data and radar data. When an accident actually occurs but the analysis results of some sensor data are inconsistent with the actual situation, the vehicle accident can still be identified based on the above-mentioned accident probability, thereby improving the accuracy of accident identification.

[0048] Furthermore, the probability of an accident is determined based on the vehicle acceleration change rate within a specified time period, the collision detection result of the environmental perception image, and the collision detection result of the radar data, including: when the vehicle collision signal is detected to be less than the collision threshold, the vehicle acceleration change rate within a specified time period before and after the vehicle collision signal is generated is obtained; a first collision confidence level is determined based on a comparison result between the vehicle acceleration change rate and the change rate threshold; an environmental perception image of the vehicle's surroundings within a specified time period before and after the vehicle collision signal is generated is obtained, and the degree of deformation of the object in front of the vehicle in the environmental perception image is detected through an image processing algorithm; a second collision confidence level is determined based on the degree of deformation of the object in front of the vehicle in the environmental perception image; and Obstacles detected by radar data within a specified time period before and after the generation of a vehicle collision signal are obtained to obtain the distance between the vehicle and the obstacle; when the distance between the vehicle and the obstacle is detected to decrease, a third collision confidence is determined according to the difference between the rate of distance decrease and the reference rate; wherein the difference is proportional to the third collision confidence; the first collision confidence is weighted by a preset first weight, the second collision confidence is weighted by a preset second weight, and the third collision confidence is weighted by a second weight, and the probability of an accident is determined by weighted summation; wherein the first weight is greater than the second weight, for example, the first weight is 0.6 and the second weight is 0.2. In order to further improve the accuracy and reduce the error, since there are differences in the accuracy of calculating the collision confidence using acceleration data, image data and radar data, a higher weight is given to the acceleration data with higher accuracy, and the same weight is given to the image data and radar data to improve the accuracy of the probability of an accident.

[0049] Among them, the first collision confidence, the second collision confidence and the third collision confidence are respectively determined by the vehicle acceleration change rate, the environmental perception image and the radar data within a specified time period. The first collision confidence is proportional to the vehicle acceleration change rate, the second collision confidence is proportional to the deformation degree of the object in front of the vehicle in the environmental perception image, and the third collision confidence is proportional to the difference between the rate at which the distance decreases and the reference rate. Compared with the method of determining whether a collision occurs, the accuracy of the detection result is improved, and then the first collision confidence, the second collision confidence, the third collision confidence and the first weight and the second weight are combined for weighted processing to obtain the probability of an accident, thereby further improving the accuracy of accident identification.

[0050] Step 104: When the accident identification result is that a vehicle accident has occurred, an accident report of the vehicle accident is generated.

[0051] In this embodiment, if the accident identification result is that a vehicle accident has occurred, then based on the time of occurrence of the vehicle accident, relevant data before and after the vehicle accident is collected, including but not limited to vehicle data such as speed, position, braking status, acceleration, steering angle, and evidence data such as images and video clips. Then, an accident report is generated based on the collected relevant data. Optionally, the accident report may include the time and location of the accident, the vehicle status (such as speed, direction), the dynamic behavior of the vehicle before and after the accident, possible causes of the accident, etc. The accident report may be in a structured format and ensure the accuracy and non-tamperability of the data. In addition, the user may be prompted to confirm when generating the accident report, and the prompting methods include but are not limited to vehicle computer display, voice prompts, etc.

[0052] According to the technical solution of the embodiment of the present disclosure, the vehicle driving data is predicted by a pre-trained recognition model to obtain an accident prediction result. Furthermore, when the accident prediction result is that there is a potential accident, the sensor data is matched with the preset accident conditions to determine the accident recognition result. Further, when the accident recognition result is that a vehicle accident has occurred, an accident report of the vehicle accident is generated. Thus, potential accidents are identified based on the vehicle driving data, and then the potential accidents are confirmed based on the sensor data. Vehicle accidents can be accurately identified. When the airbag is not triggered or the collision is greater than the threshold, vehicle accidents such as minor scratches can be identified, thereby avoiding missed vehicle accidents. When it is confirmed that a vehicle accident has occurred, an accident report is generated for accident handling, thereby improving the accuracy and efficiency of accident handling. In addition, after the potential accident is detected, it is confirmed through sensor data such as images to ensure accuracy while reducing the amount of calculation.

[0053] Based on the above embodiments, Figure 2 A flow chart of another vehicle accident handling method provided by an embodiment of the present disclosure is shown as follows: Figure 2 As shown, after generating the accident report of the vehicle accident, the method further includes the following steps:

[0054] Step 201, determining the urgency of the vehicle accident according to the vehicle state, personnel state and environmental state after the vehicle accident occurs.

[0055] In this embodiment, the vehicle status includes vehicle damage, number of accident vehicles, type of accident vehicles, etc., the personnel status includes the number of passengers, age of passengers, etc., and the environmental status includes road type, weather conditions, etc. The emergency level can be preset to multiple levels, and a corresponding relationship between the vehicle status, personnel status, and environmental status and each level is established.

[0056] As an example, the worse the vehicle condition, the higher the urgency. For example, the more serious the vehicle damage, the higher the urgency. For another example, the more the number of accident vehicles, the higher the urgency. For another example, the urgency when the accident vehicle type is a large vehicle is higher than the urgency when the accident vehicle is a non-large vehicle.

[0057] As another example, the worse the condition of the person, the higher the urgency level. For example, the more the number of passengers, the higher the urgency level. For another example, the urgency level when the age of the passengers is greater than or equal to the first age and less than or equal to the second age is lower than the urgency level when the age of the passengers is less than the first age or greater than the second age.

[0058] As another example, the worse the environmental conditions are, the higher the urgency level is. For example, the urgency level when the road type is a highway is higher than the urgency level when the road type is a non-highway, and for another example, the urgency level when the weather condition is rainy or snowy is higher than the urgency level when the weather condition is sunny.

[0059] Step 202: According to a preset accident handling strategy, an operation matching the urgency level is performed on the accident report.

[0060] In this embodiment, the generated accident report can be sent to a designated recipient through the vehicle's communication system, wherein the recipient includes an emergency response center, an insurance company, a vehicle owner, or a related service provider, etc. Optionally, an encryption algorithm is used for encryption processing when transmitting the accident report, and the encryption algorithm includes an advanced encryption standard, a public key encryption algorithm, etc.; wherein the transmission protocol can use the Hypertext Transfer Protocol Security Protocol, the Message Queue Telemetry Transfer Protocol, etc., and use the Transport Layer Security Protocol or the Secure Sockets Layer Protocol to enhance the security during data transmission; wherein, during the transmission process, the accident report can be compressed to reduce transmission time and cost.

[0061] In this embodiment, different levels of urgency may correspond to different operations. Optionally, the operations performed on the accident report include: sending the accident report to one or more of the vehicle owner's terminal, the insurance company's server, the cloud server of the Internet of Vehicles platform, and the emergency response center. Different levels of urgency may correspond to different recipients.

[0062] As an example, the degree of urgency is set from high to low as the first level, the second level, and the third level. When the degree of urgency is the first level, the accident report is sent to the car owner terminal, the insurance company server, the cloud server of the Internet of Vehicles platform, and the emergency response center at the same time. When the degree of urgency is the second level, the accident report is sent to the car owner terminal, the insurance company server, and the cloud server of the Internet of Vehicles platform. When the degree of urgency is the third level, the accident report is sent to the car owner terminal.

[0063] Optionally, accident reports of different urgency levels can be transmitted in descending order of urgency to ensure that the accident report can be transmitted first, and a retry mechanism can be designed to deal with transmission failures caused by network instability to ensure transmission stability. After the accident report is sent to the insurance company server, the insurance company server sends the evaluation results of the accident report and subsequent processing instructions to the car owner's terminal.

[0064] In the disclosed embodiment, according to the urgency of the vehicle accident, an operation matching the urgency can be performed on the accident report to adapt to vehicle accidents of different degrees.

[0065] Based on the above embodiments, Figure 3 A schematic diagram of a process for identifying fraudulent behavior in vehicle accidents provided by an embodiment of the present disclosure, such as Figure 3 As shown, after generating the accident report of the vehicle accident, the method further includes the following steps:

[0066] Step 301, obtaining vehicle internal data and external data about a vehicle accident.

[0067] Step 302 , performing fraud identification based on the vehicle internal data and external data to determine whether there is any fraudulent behavior.

[0068] In this embodiment, the vehicle's internal data includes the vehicle's historical accident information, driving behavior characteristics corresponding to the vehicle accident, and first vehicle data about the vehicle accident in the accident report; the external data includes second vehicle data about the vehicle accident collected by an external data source of the vehicle, wherein the external data source may be road traffic monitoring, checkpoint speed measurement cameras, etc. The first vehicle data and the second vehicle data include but are not limited to data such as vehicle speed.

[0069] Optionally, in order to improve the accuracy and reliability of the data, data cleaning can be performed on the data, such as data cleaning of the data from the external data source. Among them, data cleaning includes removing duplicate data, correcting erroneous data, filling missing data, standardizing data format, screening valid data, establishing data models, etc. Among them, removing duplicate data is used in the case where the same vehicle is recorded multiple times to ensure that each vehicle is not recorded repeatedly; correcting erroneous data is used in the case where data errors or damage occur during data transmission and storage, such as determining abnormal values ​​by comparing the data of the previous and next time periods, and correcting the abnormal values; filling missing data is used in the case where some data is missing, such as using interpolation to fill missing data; standardizing data formats is used to standardize data from different devices or systems, such as including data formats, units, and coding rules; screening valid data is used to retain valid information, such as setting thresholds to exclude vehicle records whose speed is not within the preset normal range, or retaining data within a specific time period, etc.

[0070] The following is an example of fraud identification.

[0071] As a possible implementation method, the first vehicle data about the vehicle accident in the accident report and the second vehicle data of the vehicle accident collected by the external data source of the vehicle are obtained, and then the first vehicle data and the second vehicle data are compared, and whether there is fraud is determined according to the comparison result. Among them, by comparing the first vehicle data and the second vehicle data, it is determined whether there is fraud according to the consistency of the comparison result. For example, taking the vehicle speed as an example, the vehicle speed obtained by the speed measuring camera at the checkpoint is compared with the actual vehicle speed within 1 minute before and after the accident time to verify whether the vehicle speed is true, and the probability of fraud is generated according to the comparison result. Optionally, the comparison result of the first vehicle data and the second vehicle data includes consistency and inconsistency, wherein the probability of fraud when the comparison result is inconsistent is greater than the probability of fraud when the comparison result is consistent, or the comparison result includes the difference between the first vehicle data and the second vehicle data, and the greater the difference, the higher the probability of fraud.

[0072] As another possible implementation method, historical accident features are extracted based on the historical accident information of the vehicle, and driving behavior features corresponding to the vehicle accident are obtained. Then, anomaly detection is performed based on the historical accident features and driving behavior features to determine whether there is fraudulent behavior. Among them, historical accident features include accident frequency, accident time, accident location, etc. Driving behavior features are used to represent driver behavior patterns. Anomaly detection is performed through historical accident features and driving behavior features to identify abnormal behavior patterns, such as frequent accidents in a short period of time, to determine the probability of fraudulent behavior. Optionally, anomaly detection is implemented using machine learning or deep learning, and cluster analysis of abnormal behavior patterns of different groups is performed to identify potential patterns of fraudulent behavior. Anomaly detection algorithms include density-based anomaly detection, isolation forest, etc. When training the model, indicators such as precision, recall, and F1 score are used to evaluate model performance, and the model is continuously optimized based on feedback and collected training data.

[0073] Optionally, based on the probability of fraudulent behavior, vehicle accidents with a probability greater than a threshold are marked to provide data support for subsequent processing. In the disclosed embodiment, fraudulent behavior related to vehicle accidents can be identified to provide data support for preventing fraudulent behavior.

[0074] Based on the above embodiments, Figure 4 A flow chart of another vehicle accident handling method provided by an embodiment of the present disclosure is shown as follows: Figure 4 As shown, the method comprises the following steps:

[0075] Step 401, obtain accident information of the vehicle accident, and send the accident information of the vehicle accident to the server.

[0076] In this embodiment, after determining that a vehicle accident has occurred, each vehicle can extract accident information based on the accident report of the vehicle accident to send the accident information of the vehicle accident to the server. The server generates accident risk information based on the accident information sent by each vehicle. The accident information includes the time of the accident, the location of the accident and the type of accident. The types of accidents include rear-end collisions, side collisions, pedestrian collisions, etc. The accident reports are classified to identify the corresponding accident types. Optionally, the potential risk factors of vehicle accidents can also be analyzed based on the accident report. The potential risk factors include weather conditions, road conditions, traffic flow, driver behavior, etc.

[0077] As an example, the server uses geographic information system technology to geocode each accident location to visualize the accident distribution on the map, identify accident-prone areas, and, based on the accident time of each vehicle accident, determine the frequency of accidents in different time periods, thereby identifying the peak time periods when accidents occur. The server generates accident risk information based on information such as accident-prone areas and peak time periods when accidents occur. The accident risk information indicates the risk of vehicle accidents occurring on different roads at different time periods.

[0078] Step 402, receiving accident risk information sent by the service end, and providing accident risk warnings during vehicle driving based on the accident risk information.

[0079] In this embodiment, after the vehicle receives the accident risk information, during the driving process of the vehicle, if it is determined that the vehicle position and the current time match the accident risk information, a prompt message is generated to provide an accident risk prompt, so that the driver can understand the driving risk and improve driving safety. Optionally, a statistical model or a machine learning model can be established to predict the accident risk. When using historical data as a training set for model training, factors such as time, weather, and traffic flow can also be combined to improve the accuracy of the prediction.

[0080] In one embodiment of the present disclosure, a vehicle driving score, an operation data factor, a road data factor, a driving data factor, and a business data factor can be extracted based on an accident report, and a recommendation model can be established based on the vehicle driving score, the operation data factor, the road data factor, the driving data factor, and the business data factor to generate recommendation information, wherein the recommendation information can be used for driving advice prompts, premium assessment, etc. Among them, the vehicle driving score model is used to generate a driver's driving score based on policy data, claim data, driving license data, vehicle data, historical driving data, etc., combined with weights, and the driving score can adopt multiple levels. The operation data factor is used to evaluate the vehicle's operating efficiency, cost control, and maintenance requirements based on the vehicle's mileage, fuel consumption, maintenance records, and failure frequency, etc., to provide data support for optimizing vehicle use plans, etc. The road data factor evaluates the safety of vehicle driving based on road type, road conditions, traffic flow, weather conditions, etc., so as to plan driving routes and reduce the probability of accidents, etc. Driving data factors include vehicle speed, acceleration, braking frequency, steering angle, etc. By analyzing driving data factors, information such as the driver's driving style is generated to provide data support for the formulation of safety improvement plans. Business data factors include order volume, delivery time, customer satisfaction, income and expenditure, etc. By analyzing business data factors, data support is provided for enterprises to optimize business processes.

[0081] Figure 5 A schematic diagram of the structure of a vehicle accident handling device provided by an embodiment of the present disclosure, such as Figure 5 As shown, the vehicle accident handling device includes: an acquisition module 51, a prediction module 52, a determination module 53, and a generation module 54.

[0082] Wherein, the acquisition module 51 is used to acquire vehicle driving data;

[0083] The prediction module 52 is used to perform accident prediction on the vehicle driving data based on the pre-trained recognition model to obtain an accident prediction result; the accident prediction result indicates whether there is a potential accident;

[0084] The determination module 53 is used to obtain sensor data when the accident prediction result indicates that there is a potential accident, and match the sensor data with the preset accident conditions to determine the accident identification result;

[0085] The generating module 54 is used to generate an accident report of the vehicle accident when the accident identification result is that a vehicle accident has occurred.

[0086] In one embodiment of the present disclosure, the acquisition module 51 is specifically used to:

[0087] Acquiring vehicle driving data at a first acquisition frequency according to a preset period;

[0088] and / or,

[0089] In response to detecting an abnormal vehicle driving event, vehicle driving data is acquired at a second acquisition frequency during a preset time period; wherein the second acquisition frequency is greater than the first acquisition frequency.

[0090] In one embodiment of the present disclosure, the determination module 53 is specifically used to:

[0091] When the vehicle collision signal is detected to be greater than or equal to the collision threshold, the accident recognition result is determined to be a vehicle accident;

[0092] and / or,

[0093] When a rescue signal is detected from a vehicle, the accident identification result is determined to be a vehicle accident;

[0094] and / or,

[0095] Determine the probability of an accident based on the rate of change of vehicle acceleration within a specified time period, the collision detection results of the environmental perception image, and the collision detection results of the radar data;

[0096] If the probability of an accident occurring is greater than a preset threshold, the accident identification result is determined to be a vehicle accident occurring.

[0097] In one embodiment of the present disclosure, the device further includes:

[0098] A training module, used to obtain training data; the training data includes first sample driving data within a preset time period before and after the accident, and second sample driving data within a preset time period without the accident;

[0099] The first sample driving data is used as a positive sample and the second sample driving data is used as a negative sample to train an initial recognition model to generate the pre-trained recognition model.

[0100] In one embodiment of the present disclosure, the device further includes:

[0101] A transmission module, used to determine the urgency of the vehicle accident according to the vehicle state, personnel state and environmental state after the vehicle accident occurs;

[0102] Based on the preset incident handling strategy, perform actions that match the urgency of the incident report.

[0103] In one embodiment of the present disclosure, the device further includes:

[0104] An identification module, used to obtain first vehicle data about the vehicle accident in the accident report and second vehicle data about the vehicle accident collected by an external data source of the vehicle;

[0105] Comparing the first vehicle data with the second vehicle data, and determining whether there is fraudulent behavior according to the comparison result; and / or,

[0106] Extract historical accident features based on historical accident information of the vehicle, and obtain driving behavior features corresponding to the vehicle accident;

[0107] Anomaly detection is performed based on historical accident characteristics and driving behavior characteristics to determine whether there is fraudulent behavior.

[0108] In one embodiment of the present disclosure, the device further includes:

[0109] A prompt module is used to extract accident information of vehicle accidents based on accident reports and send the accident information of vehicle accidents to a server; wherein the accident information includes the time of the accident, the location of the accident and the type of the accident. The server is used to generate accident risk information according to the accident information sent by each vehicle;

[0110] Receive accident risk information sent by the server, and provide accident risk warnings during vehicle driving based on the accident risk information.

[0111] In one embodiment of the present disclosure, the determination module 53 is specifically used to:

[0112] When it is detected that the vehicle collision signal is less than the collision threshold, obtaining the vehicle acceleration change rate within a specified time period before and after the vehicle collision signal is generated;

[0113] Determining a first collision confidence level according to a comparison result between the vehicle acceleration change rate and a change rate threshold; wherein the vehicle acceleration change rate is proportional to the first collision confidence level;

[0114] Acquire an environmental perception image of the vehicle's surroundings within a specified time period before and after the vehicle collision signal is generated, and detect the degree of deformation of the object in front of the vehicle in the environmental perception image through an image processing algorithm;

[0115] Determining a second collision confidence level according to a deformation degree of the object in front of the vehicle in the environment perception image; wherein the deformation degree is proportional to the second collision confidence level;

[0116] For an obstacle detected by the radar data within a specified time period before and after a vehicle collision signal is generated, obtaining a distance between the vehicle and the obstacle;

[0117] When it is detected that the distance between the vehicle and the obstacle decreases, a third collision confidence is determined according to a difference between a rate at which the distance decreases and a reference rate; wherein the difference is proportional to the third collision confidence;

[0118] The first collision confidence is weighted using a preset first weight, the second collision confidence is weighted using a preset second weight, and the third collision confidence is weighted using the second weight, and the accident probability is determined by weighted summation; wherein the first weight is greater than the second weight.

[0119] The vehicle accident handling device provided in the embodiment of the present disclosure can execute any vehicle accident handling method provided in the embodiment of the present disclosure, and has the corresponding functional modules and beneficial effects of the execution method. The contents not described in detail in the embodiment of the device of the present disclosure can refer to the description in any method embodiment of the present disclosure.

[0120] The present disclosure also provides an electronic device, which includes one or more processors and a memory. The processor may be a central processing unit (CPU) or other forms of processing units with data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device to perform desired functions. The memory may include one or more computer program products, and the computer program product may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. One or more computer program instructions may be stored on a computer-readable storage medium, and the processor may run the program instructions to implement the method of the above embodiment of the present disclosure and / or other desired functions. Various contents such as input signals, signal components, noise components, etc. may also be stored in the computer-readable storage medium.

[0121] In one example, the electronic device may further include: an input device and an output device, and these components are interconnected through a bus system and / or other forms of connection mechanisms. In addition, the input device may also include, for example, a keyboard, a mouse, etc. The output device may output various information to the outside, including determined distance information, direction information, etc. The output device may include, for example, a display, a speaker, a printer, and a communication network and a remote output device connected thereto, etc. In addition, according to the specific application, the electronic device may also include any other appropriate components such as a bus, an input / output interface, etc.

[0122] In addition to the above methods and devices, the embodiments of the present disclosure may also be a computer program product, which includes computer program instructions. When the computer program instructions are executed by a processor, the processor executes any method provided by the embodiments of the present disclosure.

[0123] The computer program product may be written in any combination of one or more programming languages ​​to write program code for performing the operations of the disclosed embodiments, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0124] In addition, the embodiments of the present disclosure may also be a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the processor executes any method provided by the embodiments of the present disclosure.

[0125] Computer readable storage media can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can include, for example, but is not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0126] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0127] The above description is only a specific embodiment of the present disclosure, so that those skilled in the art can understand or implement the present disclosure. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure will not be limited to the embodiments described herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for handling a vehicle accident, characterized in that: The method comprises: Obtain vehicle driving data; Performing accident prediction on the vehicle driving data based on a pre-trained recognition model to obtain an accident prediction result; the accident prediction result indicates whether there is a potential accident; When the accident prediction result indicates that there is a potential accident, obtaining sensor data, and matching the sensor data with a preset accident condition to determine an accident identification result; When the accident identification result is that a vehicle accident has occurred, an accident report of the vehicle accident is generated.

2. The method according to claim 1, characterized in that The obtaining of vehicle driving data comprises: Acquiring vehicle driving data at a first acquisition frequency according to a preset period; and / or, In response to detecting an abnormal vehicle driving event, vehicle driving data is acquired at a second acquisition frequency during a preset time period; wherein the second acquisition frequency is greater than the first acquisition frequency.

3. The method according to claim 1, characterized in that The matching of the sensor data with a preset accident condition to determine an accident identification result includes: When it is detected that the vehicle collision signal is greater than or equal to the collision threshold, determining that the accident recognition result is a vehicle accident; and / or, When a rescue signal is detected from the vehicle, determining that the accident identification result is a vehicle accident; and / or, Determine the probability of an accident based on the rate of change of vehicle acceleration within a specified time period, the collision detection results of the environmental perception image, and the collision detection results of the radar data; If the probability of the accident occurring is greater than a preset threshold, the accident identification result is determined to be a vehicle accident occurring.

4. The method according to claim 1, characterized in that The method further comprises: Acquire training data; the training data includes first sample driving data within a preset time period before and after the accident, and second sample driving data within a preset time period when no accident occurs; The first sample driving data is used as a positive sample and the second sample driving data is used as a negative sample to train an initial recognition model to generate the pre-trained recognition model.

5. The method according to claim 1, characterized in that After generating an accident report for the vehicle accident, the method further includes: Acquire first vehicle data about the vehicle accident in the accident report and second vehicle data about the vehicle accident collected by an external data source of the vehicle; Compare the first vehicle data with the second vehicle data, and determine whether there is fraud according to the comparison result; and / or, Extracting historical accident features based on historical accident information of the vehicle, and obtaining driving behavior features corresponding to the vehicle accident; Anomaly detection is performed based on the historical accident characteristics and the driving behavior characteristics to determine whether the fraudulent behavior exists.

6. The method according to claim 1, characterized in that After generating an accident report for the vehicle accident, the method further includes: Extracting accident information of the vehicle accident based on the accident report, and sending the accident information of the vehicle accident to a server; wherein the accident information includes the time of the accident, the location of the accident and the type of the accident, and the server is used to generate accident risk information according to the accident information sent by each vehicle; Receive the accident risk information sent by the service end, and provide accident risk warnings during vehicle driving based on the accident risk information.

7. The method according to claim 3, characterized in that The determining of the probability of an accident according to the vehicle acceleration change rate within a specified time period, the collision detection result of the environment perception image, and the collision detection result of the radar data includes: Determining a first collision confidence level according to a comparison result between the vehicle acceleration change rate and the change rate threshold; wherein the vehicle acceleration change rate is proportional to the first collision confidence level; Determining a second collision confidence level according to a deformation degree of an object in front of the vehicle in the environment perception image; wherein the deformation degree is proportional to the second collision confidence level; When it is detected that the distance between the vehicle and the obstacle detected by the radar data decreases, a third collision confidence is determined according to a difference between a rate at which the distance decreases and a reference rate; wherein the difference is proportional to the third collision confidence; The first collision confidence is weighted using a preset first weight, the second collision confidence is weighted using a preset second weight, and the third collision confidence is weighted using the second weight, and the accident probability is determined by weighted summation; wherein the first weight is greater than the second weight.

8. A vehicle accident handling device, characterized in that: include: An acquisition module, used to acquire vehicle driving data; A prediction module, used to perform accident prediction on the vehicle driving data based on a pre-trained recognition model to obtain an accident prediction result; the accident prediction result indicates whether there is a potential accident; A determination module, used for obtaining sensor data when the accident prediction result indicates that a potential accident exists, and matching the sensor data with a preset accident condition to determine an accident identification result; A generating module is used to generate an accident report of the vehicle accident when the accident identification result is that a vehicle accident has occurred.

9. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is used to read the executable instructions from the memory and execute the instructions to implement the vehicle accident handling method described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the vehicle accident handling method described in any one of claims 1 to 7 is implemented.

Images