A method for processing vehicle accident verification

Abstract

The present application relates to the field of vehicle accident verification technology, and in particular to a method for processing vehicle accident verification. The method comprises: in response to a verification request, obtaining a vehicle operation status data set that matches the unique identifier of the target vehicle and the date of the accident from a connected vehicle data platform; if the data set matches the location of the accident included in the verification request, determining whether the vehicle operation status data of a first preset type included in the data set meets an initial preset judgment condition; if not, determining a first probability of the target vehicle having an accident at each parking time node based on the vehicle operation status data of a second preset type corresponding to the target time period of each parking time node in the data set; if p 0,1 <p i <p 0,2 , then the third preset type of data corresponding to the i-th parking time node is used to determine whether an accident occurred at the parking time point. The present invention can improve the accuracy of accident verification of connected vehicles.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle accident verification, and in particular to a processing method for vehicle accident verification. Background Art

[0002] When a vehicle is involved in an accident, it is necessary to verify the accident. Currently, there are generally the following methods for verifying the accident:

[0003] The first is to provide post-incident evidence collection, typically requiring the vehicle owner or surveyor to take photos and videos of the vehicle. The time and location of the photos are then used to verify the time and location of the accident. Image recognition algorithms are then used to identify the damaged parts of the vehicle and estimate the amount of damage, thereby verifying the authenticity of the accident. This method has led to cases where the vehicle owner has lied about the accident, fabricated the scene, or even colluded with surveyors to fabricate a report.

[0004] The second approach is to use pre-built system functions or external devices to promptly identify a collision through sensors and collision recognition algorithms. After the collision, evidence can be collected by taking photos, videos, and recording various vehicle status data. This approach requires a collision recognition algorithm or external devices, which is costly and has low coverage.

[0005] Currently, connected vehicles in China generally have a vehicle status data reporting mechanism, which typically reports basic vehicle information (model, vehicle number, etc.) and vehicle operating status data to the connected vehicle data platform. Vehicle operating status data includes speed, acceleration, angular velocity, geographic location, seatbelt status, headlight status, vehicle gear position, door switch status, brake and throttle operation, etc. This data is reported at a fixed frequency while the vehicle is in motion, and most connected vehicles have this reporting mechanism, with a high coverage rate. How to improve the accuracy of connected vehicle accident verification is an urgent problem that needs to be solved. Summary of the Invention

[0006] The purpose of the present invention is to provide a method for processing vehicle accident verification to improve the accuracy of accident verification of connected vehicles.

[0007] According to the present invention, a method for processing vehicle accident verification is provided, the method comprising the following steps:

[0008] S100, in response to a verification request, obtaining a vehicle operating status data set that matches the unique identifier of the target vehicle and the date of the accident from the connected vehicle data platform; the verification request includes the unique identifier of the target vehicle, the date of the accident, and the location of the accident.

[0009] S200, determine whether the vehicle operating status data set matches the location of the accident included in the verification request. If so, enter S300.

[0010] S300, judging whether the vehicle operating status data of the first preset type included in the vehicle operating status data set meets the initial preset judgment condition, if so, judging that the verification is passed; otherwise, entering S400.

[0011] S400, obtaining each parking time node of the target vehicle within the date of the accident occurrence according to the preset vehicle parking judgment condition, and judging the first probability of the target vehicle having an accident at each parking time node according to the second preset type of vehicle operation status data of the target time period corresponding to each parking time node in the vehicle operation status data set.

[0012] S500, if p 0,1 <p i <p 0,2 , then enter S600; p 0,1 and p 0,2 are the first preset probability and the second preset probability respectively, p i is the first probability of the target vehicle having an accident at the i-th parking time point.

[0013] S600, judging whether the target vehicle has an accident at the i-th parking time point based on the vehicle operation status data of the third preset type corresponding to the i-th parking time node in the vehicle operation status data set; if the target vehicle has no accident at any parking time point, the verification is determined to have failed; otherwise, the verification is determined to have passed.

[0014] Compared with the prior art, the present invention has at least the following beneficial effects:

[0015] The present invention first obtains the vehicle operation status data set of the corresponding date of the target vehicle according to the verification request, and performs multi-level judgment based on the vehicle operation status data set, including verifying whether the vehicle's location data matches the location of the accident. If it matches, the vehicle operation status data of different preset types are processed differently in turn. Among them, the process of judging according to the first preset type of vehicle operation status data is simple, and the verification result can be quickly obtained when the vehicle operation status data of the first preset type meets the initial preset judgment conditions; if the initial preset judgment conditions are not met, each parking time node is judged and analyzed, and the verification is judged to be unsuccessful only when the target vehicle has no accident at any parking time point; and when judging and analyzing each parking time node, the vehicle operation status data of different preset types are processed differently and comprehensively judged according to the vehicle operation status data of different preset types; thereby, the present invention achieves the purpose of quickly and accurately verifying accidents involving connected vehicles. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 This is a flowchart of a method for processing vehicle accident verification provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0019] According to this embodiment, Figure 1 As shown, a method for processing vehicle accident verification is provided, the method comprising the following steps:

[0020] S100, in response to a verification request, obtaining a vehicle operating status data set that matches the unique identifier of the target vehicle and the date of the accident from the connected vehicle data platform; the verification request includes the unique identifier of the target vehicle, the date of the accident, and the location of the accident.

[0021] In this embodiment, the connected vehicle data platform stores the unique identifier of each connected vehicle, the data date, and the corresponding vehicle operating status dataset. By comparing the target vehicle's unique identifier and accident date with the unique identifiers and data dates of each connected vehicle stored on the connected vehicle data platform, a vehicle operating status dataset that matches the target vehicle's unique identifier and accident date can be obtained. Optionally, the vehicle's unique identifier is the vehicle's license plate number or vehicle frame number.

[0022] S200, determine whether the vehicle operating status data set matches the location of the accident included in the verification request. If so, enter S300.

[0023] In this embodiment, S200 further includes: if there is no match, determining that the verification fails.

[0024] As a specific embodiment, the vehicle operating status data set includes vehicle location data, and S200 includes:

[0025] S210, obtaining vehicle position data from the vehicle operation status data set according to a preset time interval; wherein the time interval corresponding to any two adjacent vehicle position data is the preset time interval.

[0026] In this embodiment, the vehicle operation status data set includes the vehicle position data at each collection moment, and the vehicle position data is extracted according to a preset time interval. The preset time interval is an empirical value. Optionally, the preset time interval is 30 minutes or 1 hour.

[0027] As a specific implementation, the vehicle location data is Global Positioning System (GPS) data.

[0028] S220, if any of the acquired vehicle location data does not belong to the same district or county as the location where the accident occurred, it is determined that the vehicle operation status dataset does not match the location where the accident occurred included in the verification request; otherwise, it is determined that the vehicle operation status dataset matches the location where the accident occurred included in the verification request.

[0029] Based on S210-S220, a preliminary and rapid verification of the location where the vehicle accident occurred can be achieved.

[0030] S300, judging whether the vehicle operating status data of the first preset type included in the vehicle operating status data set meets the initial preset judgment condition, if so, judging that the verification is passed; otherwise, entering S400.

[0031] In this embodiment, the first preset type of vehicle operating state is known, and the first preset type of vehicle operating state is the vehicle operating state with the greatest correlation with the vehicle accident, which can be obtained based on experience or the Pearson correlation coefficient; for example, the first preset type of vehicle operating state is the state of the airbag, and the initial preset judgment condition is that the airbag is opened.

[0032] S400, obtaining each parking time node of the target vehicle within the date of the accident occurrence according to the preset vehicle parking judgment condition, and judging the first probability of the target vehicle having an accident at each parking time node according to the second preset type of vehicle operation status data of the target time period corresponding to each parking time node in the vehicle operation status data set.

[0033] In this embodiment, the preset vehicle parking determination conditions are determined based on experience. Optionally, the preset vehicle parking determination conditions include: the vehicle speed is zero and the duration is greater than or equal to a preset time threshold, or no vehicle operating status data is reported after the vehicle speed reaches zero. Optionally, the preset time threshold is 10 minutes.

[0034] In this embodiment, the second preset type of vehicle operating state is known and can be determined empirically or using a Pearson correlation coefficient, such as vehicle speed, acceleration, and steering angle. The correlation between the second preset type of vehicle operating state and the occurrence of a vehicle accident is less than the correlation between the first preset type of vehicle operating state and the occurrence of a vehicle accident.

[0035] Optionally, a machine learning or deep learning model (such as a neural network model, a support vector machine, or a random forest) is used to determine the first probability of the target vehicle having an accident at each parking time node. As a preferred embodiment, S400 includes:

[0036] S410, obtain the second preset type of vehicle operation status data A of the target time period corresponding to the i-th parking time node; the upper limit of the target time period corresponding to the i-th parking time node is the moment that is the first preset time length away from the i-th parking time node and earlier than the i-th parking time node, and the lower limit of the target time period corresponding to the i-th parking time node is the moment that is the first preset time length away from the i-th parking time node and later than the i-th parking time node.

[0037] In this embodiment, the first preset time length is an empirical value. Optionally, the first preset time length is 10 seconds.

[0038] S420, construct the feature vector F of the i-th parking time node according to A; F=(F1, F2, ..., F j ,…,F m ), F jis a data sequence of the jth vehicle operating state of the second preset type, where j ranges from 1 to m, and m is the number of vehicle operating states of the second preset type; F j =(F j,1 ,F j,2 ,…,F j,r ,…,F j,R ), F j,r The data of the rth collection moment of the jth vehicle operating state of the second preset type within the target time period corresponding to the i-th parking time node is collected. The value range of r is 1 to R, and R is the number of collection moments included in the target time period corresponding to the i-th parking time node.

[0039] S430, input the feature vector of the i-th parking time node into the trained target neural network model, and determine the output of the target neural network model as the first probability of the target vehicle having an accident at the i-th parking time node; the trained target neural network model has the function of inferring the probability of an accident based on the input feature vector.

[0040] Those skilled in the art will appreciate that any prior art training method for a neural network model falls within the scope of protection of the present invention. Optionally, the target neural network model is a multi-layer perceptron (MLP).

[0041] In this embodiment, the second preset type of vehicle operating status is a vehicle operating status that is highly correlated with a vehicle accident. Based on S410-S430, the probability of a vehicle accident can be obtained based on the second preset type of vehicle operating status data, and based on this probability, a quick judgment can be made as to whether the vehicle has an accident.

[0042] S500, if p 0,1 <p i <p 0,2 , then enter S600; p 0,1 and p 0,2 are the first preset probability and the second preset probability respectively, p i is the first probability of the target vehicle having an accident at the i-th parking time point.

[0043] In this embodiment, S500 further includes: if p i ≥p 0,2 , then it is determined that the target vehicle has an accident at the i-th parking time point and the verification is passed. i ≤p 0,1 , it is determined that the target vehicle has no accident at the i-th parking time point; if the target vehicle has no accident at any parking time point, it is determined that the verification has failed; otherwise, it is determined that the verification has passed.

[0044] In this embodiment, the value range of i is 1 to n, where n is the number of parking time nodes of the target vehicle on the date when the accident occurs.

[0045] In this embodiment, p 0,1 and p 0,2 All are experience values, optional, p 0,1 =0.2, p 0,2 =0.8.

[0046] S600, judging whether the target vehicle has an accident at the i-th parking time point based on the vehicle operation status data of the third preset type corresponding to the i-th parking time node in the vehicle operation status data set; if the target vehicle has no accident at any parking time point, the verification is determined to have failed; otherwise, the verification is determined to have passed.

[0047] As a preferred embodiment, S600 includes:

[0048] S610: Determine the second probability f of the target vehicle having an accident at the i-th parking time point based on the third preset type of vehicle operation status data corresponding to the i-th parking time node in the vehicle operation status data set. i ; The correlation coefficient between the vehicle operating status data of the third preset type and the vehicle accident is less than the correlation coefficient between the vehicle operating status data of the second preset type and the vehicle accident.

[0049] In this embodiment, the third preset type of vehicle operating state is known and can be determined based on experience or using the Pearson correlation coefficient, such as the working state of the double flash lights and the working state of the turn signal lights. Compared with the second preset type of vehicle operating state, the third preset type of vehicle operating state has a relatively small correlation coefficient with the occurrence of a vehicle accident. However, the third preset type of vehicle operating state is also associated with the occurrence of a vehicle accident. In this embodiment, the third preset type of vehicle operating state is used for p 0,1 <p i <p 0,2 Assisted judgment in the case.

[0050] As a preferred embodiment, S610 includes:

[0051] S611, obtaining the vehicle running state data B of the third preset type corresponding to the i-th parking time node; B=(B1, B2, ..., B x ,…,B y ), B x The data of the xth vehicle operating state of the third preset type in the xth time period corresponding to the i-th parking time node is x, where the value of x ranges from 1 to y, and y is the number of vehicle operating states of the third preset type.

[0052] In this embodiment, the duration and start time of the xth time period are empirically determined; different xth time periods may have the same or different durations and start times. For example, if the first third preset type of vehicle operating state is the hazard lights operating state, the first time period is within 5 minutes after the parking time point; if the second third preset type of vehicle operating state is the turn signal operating state, the second time period is within 5 minutes after the parking time point.

[0053] S612, traverse B, judge B x Whether the xth preset condition is met.

[0054] In this embodiment, the xth preset condition represents a condition that is typically satisfied by the xth third preset type of vehicle operating state during the xth time period corresponding to the parking time node when the vehicle is involved in an accident. The xth preset condition can be set in advance based on experience. For example, if the first third preset type of vehicle operating state is the hazard lights operating state, the first preset condition is that the hazard lights are activated within 5 minutes of the parking time node; if the second third preset type of vehicle operating state is the turn signal operating state, the second preset condition is that the turn signal is activated within 5 minutes of the parking time node.

[0055] S613, B that meets the corresponding preset conditions in B x The ratio of the number of y is determined as f i .

[0056] In this embodiment, f i The larger the value of , the greater the probability that the target vehicle will have an accident at the i-th parking time node.

[0057] S620, according to f i P i Make adjustments and use the adjustment result p' i Determine whether the target vehicle has an accident at the i-th parking time point; p' i =k×f i +p i , k is the preset adjustment coefficient, 0 <k≤1-p 0,2 .

[0058] In this embodiment, f i With p' i Positive correlation.

[0059] In this embodiment, S620 further includes: if p' i ≥p 0,2 , it is determined that the target vehicle has an accident at the i-th parking time point; otherwise, it is determined that the target vehicle has not an accident at the i-th parking time point.

[0060] As a preferred embodiment, the process of obtaining k includes:

[0061] S621, obtaining a sample vehicle data set; the sample vehicle data set includes vehicle operation status data of a second preset type and corresponding vehicle operation status data of a third preset type for a target time period corresponding to a parking time point of several sample vehicles.

[0062] S622, obtain 1-p according to the sample vehicle data set 0,2 Determine the accuracy c0 when k is used.

[0063] As a specific embodiment, the first probability of each sample vehicle having an accident at the corresponding parking time point is obtained based on the vehicle operation status data of the second preset type corresponding to the target time period at each sample vehicle at a parking time point included in the sample vehicle data set; if the first probability of a sample vehicle having an accident at the corresponding parking time point is greater than p 0,1 and is less than p 0,2 , the sample vehicle is determined to be a first type of vehicle, and the second probability of the sample vehicle having an accident at the corresponding parking time point is determined based on the vehicle operation status data of the third preset type corresponding to the sample vehicle at the corresponding parking time node. The first probability corresponding to the sample vehicle is adjusted based on the second probability of the sample vehicle having an accident at the corresponding parking time point, and the value of k during the adjustment is 1-p 0,2 , and judge whether an accident occurs based on the adjustment results; the ratio of the number of correct judgments based on the adjustment results to the number of vehicles in the first category is determined as the accuracy rate c0.

[0064] S623: If c0 is less than a preset accuracy threshold, initialize the first variable e to 1.

[0065] In this embodiment, the preset accuracy threshold is an empirical value. Optionally, the preset accuracy threshold is 0.8.

[0066] S624, obtain 1-p 0,2 -e×d determines the accuracy c when k e ; d is the preset adjustment step.

[0067] In this embodiment, d is an empirical value; optionally, d is 0.01×(1-p 0,2 ).

[0068] In this embodiment, 1-p 0,2 -e×d determines the accuracy c when k e The process is the same as above to obtain 1-p 0,2 The process of determining the accuracy c0 when k is similar and will not be repeated here.

[0069] S625, if c e If the accuracy is less than the preset threshold, then update e to e+1 and repeat S624 until c e Greater than or equal to the preset accuracy threshold, and 1-p 0,2 -e×d is determined to be k.

[0070] In this embodiment, k obtained based on S621-S625 can improve the accuracy of vehicle accident verification.

[0071] In this embodiment, 0,1 <p i <p 0,2 In this case, it is possible to accurately assist in judging whether an accident occurs to the target vehicle at the parking time node based on the vehicle operation status data of the third preset type, thereby improving the accuracy of subsequent judgment results.

[0072] This embodiment first obtains the vehicle operation status data set of the corresponding date of the target vehicle according to the verification request, and performs multi-level judgment based on the vehicle operation status data set, including verifying whether the vehicle's location data matches the location of the accident. If it matches, the vehicle operation status data of different preset types are processed differently in turn. Among them, the process of judging according to the vehicle operation status data of the first preset type is simple, and the result of verification passing can be quickly obtained when the vehicle operation status data of the first preset type meets the initial preset judgment conditions; if the initial preset judgment conditions are not met, each parking time node is judged and analyzed, and the verification is judged to be failed only when the target vehicle has not had an accident at any parking time point; and when judging and analyzing each parking time node, the vehicle operation status data of different preset types are processed differently and comprehensively judged according to the different correlation coefficients between the vehicle operation status data of different preset types and the vehicle accident; thus, this embodiment achieves the purpose of quickly and accurately verifying the accident of the connected vehicle.

[0073] Although some specific embodiments of the present invention have been described in detail by way of example, it should be understood by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present invention. It should also be understood by those skilled in the art that various modifications may be made to the embodiments without departing from the scope and spirit of the present invention. The scope of the present invention is defined by the appended claims.

Claims

1. A method for processing vehicle accident verification, characterized in that: The method comprises the following steps: S100, in response to a verification request, obtaining a vehicle operating status dataset that matches the unique identifier of the target vehicle and the date of the accident from a connected vehicle data platform; the verification request includes the unique identifier of the target vehicle, the date of the accident, and the location of the accident; S200, determining whether the vehicle operating status dataset matches the location of the accident included in the verification request, if so, proceeding to S300; S300, determining whether the first preset type of vehicle operating status data included in the vehicle operating status data set meets the initial preset judgment condition. If so, the verification is determined to be passed; otherwise, proceeding to S400; S400, obtaining each parking time node of the target vehicle within the date of the accident occurrence according to a preset vehicle parking judgment condition, and determining a first probability of the target vehicle having an accident at each parking time node based on vehicle operation status data of a second preset type corresponding to a target time period for each parking time node in the vehicle operation status dataset; S500, if p 0,1 <p i <p 0,2 , then enter S600; p 0,1 and p 0,2 are the first preset probability and the second preset probability respectively, p i is the first probability of the target vehicle having an accident at the i-th parking time point; S600, determining whether the target vehicle has an accident at the i-th parking time point based on the vehicle operation status data of a third preset type corresponding to the i-th parking time point in the vehicle operation status data set; if the target vehicle has not an accident at any parking time point, determining that the verification has failed; otherwise, determining that the verification has passed; S600 includes: S610: Determine the second probability f of the target vehicle having an accident at the i-th parking time point based on the third preset type of vehicle operation status data corresponding to the i-th parking time node in the vehicle operation status data set. i The correlation coefficient between the vehicle operation status data of the third preset type and the vehicle accident is less than the correlation coefficient between the vehicle operation status data of the second preset type and the vehicle accident; S620, according to f i P i Make adjustments and use the adjustment result p' i Determine whether the target vehicle has an accident at the i-th parking time point; p' i =k×f i +p i , k is the preset adjustment coefficient, 0 <k≤1-p 0,2 .

2. The vehicle accident verification processing method according to claim 1, characterized in that: The vehicle operation status data set includes vehicle location data, and S200 includes: S210, obtaining vehicle position data from the vehicle operation status data set according to a preset time interval; wherein the time interval corresponding to any two adjacent vehicle position data is the preset time interval; S220, if any of the acquired vehicle location data does not belong to the same district or county as the location where the accident occurred, it is determined that the vehicle operation status dataset does not match the location where the accident occurred included in the verification request; otherwise, it is determined that the vehicle operation status dataset matches the location where the accident occurred included in the verification request.

3. The vehicle accident verification processing method according to claim 1, characterized in that: The process of obtaining k includes: S621, obtaining a sample vehicle data set; the sample vehicle data set includes vehicle operation status data of a second preset type and corresponding vehicle operation status data of a third preset type for a target time period corresponding to a parking time point of a plurality of sample vehicles; S622, obtain 1-p according to the sample vehicle data set 0,2 Determine the accuracy c0 when k is set; S623, if c0 is less than the preset accuracy threshold, initialize the first variable e to 1; S624, obtain 1-p 0,2 -e×d determines the accuracy c when k e ; d is the preset adjustment step; S625, if c e If the accuracy is less than the preset threshold, then update e to e+1 and repeat S624 until c e Greater than or equal to the preset accuracy threshold, and 1-p 0,2 -e×d is determined to be k.

4. The vehicle accident verification processing method according to claim 1, characterized in that: The S610 includes: S611, obtaining the vehicle running status data B of the third preset type corresponding to the i-th parking time node; B=(B1, B2, ..., B x ,…,B y ), B x The data of the xth vehicle operating state of the third preset type in the xth time period corresponding to the i-th parking time node, where x ranges from 1 to y, and y is the number of vehicle operating states of the third preset type; S612, traverse B, judge B x Whether the xth preset condition is met; S613, B that meets the corresponding preset conditions in B x The ratio of the number of y is determined as f i .

5. The vehicle accident verification processing method according to claim 1, characterized in that: S400 includes: S410, obtaining vehicle operating status data A of a second preset type for a target time period corresponding to the i-th parking time node; the upper limit of the target time period corresponding to the i-th parking time node is a time that is a first preset time period away from the i-th parking time node and earlier than the i-th parking time node, and the lower limit of the target time period corresponding to the i-th parking time node is a time that is a first preset time period away from the i-th parking time node and later than the i-th parking time node; S420, construct the feature vector F of the i-th parking time node according to A; F=(F1, F2, ..., F j ,…,F m ), F j is a data sequence of the jth vehicle operating state of the second preset type, where j ranges from 1 to m, and m is the number of vehicle operating states of the second preset type; F j =(F j,1 ,F j,2 ,…,F j,r ,…,F j,R ), F j,r The data of the j-th vehicle operating state of the second preset type at the r-th collection moment within the target time period corresponding to the i-th parking time node is collected. The value of r ranges from 1 to R, and R is the number of collection moments included in the target time period corresponding to the i-th parking time node. S430, input the feature vector of the i-th parking time node into the trained target neural network model, and determine the output of the target neural network model as the first probability of the target vehicle having an accident at the i-th parking time node; the trained target neural network model has the function of inferring the probability of an accident based on the input feature vector.

6. The vehicle accident verification processing method according to claim 1, characterized in that: The preset vehicle parking judgment conditions include: the vehicle speed is 0 and the duration is greater than or equal to a preset time threshold, or there is no more vehicle running status data reported after the vehicle speed reaches 0.

7. The vehicle accident verification processing method according to claim 1, characterized in that: S620 further includes: if p' i ≥p 0,2 , it is determined that the target vehicle has an accident at the i-th parking time point; otherwise, it is determined that the target vehicle has not an accident at the i-th parking time point.

8. The vehicle accident verification processing method according to claim 1, characterized in that: S200 also includes: if there is no match, determining that the verification fails.

9. The vehicle accident verification processing method according to claim 1, characterized in that: S500 also includes: If p i ≥p 0,2 , then it is determined that the target vehicle has an accident at the i-th parking time point and the verification is passed; if p i ≤p 0,1 , it is determined that the target vehicle has no accident at the i-th parking time point.

Images