WO2020215690A1

WO2020215690A1 - Driving data analysis method and apparatus, and computer device and storage medium

Info

Publication number: WO2020215690A1
Application number: PCT/CN2019/118075
Authority: WO
Inventors: 李红伟
Original assignee: 平安科技（深圳）有限公司
Priority date: 2019-04-23
Filing date: 2019-11-13
Publication date: 2020-10-29
Also published as: CN110197119B; CN110197119A

Abstract

A driving data analysis method based on big data, comprising: obtaining driving data of a target vehicle, the driving data comprising a driving image; determining an identification area in the driving image; identifying a nearby vehicle with a vehicle identifier appearing in the identification area, and recording the vehicle position of the nearby vehicle; determining whether the target vehicle has an overtaking behavior or not by comparing changes in the vehicle positions of nearby vehicles in a plurality of adjacent frames of driving images; collecting statistics about the overtaking frequency of the target vehicle according to the determining result; and calculating the vehicle insurance cost corresponding to the target vehicle according to the overtaking frequency.

Description

Driving data analysis method, device, computer equipment and storage medium

This application claims the priority of a Chinese patent application filed with the Chinese Patent Office on April 23, 2019. The application number is 201910326600X, and the application name is "Driving Data Analysis Method, Device, Computer Equipment, and Storage Medium". The entire content is by reference Incorporated in this application.

Technical field

This application relates to a driving data analysis method, device, computer equipment and storage medium.

Background technique

As automobiles have gradually become a common means of transportation, the auto insurance market has also developed rapidly, and the auto insurance business has shown a significant increase. In order to promote the development of auto insurance business, a UBI (Usage Based Insurance) insurance is emerging. UBI can adjust insurance premiums based on driving data. In theory, users with safer driving behavior should receive premium discounts. However, the driving behavior data of vehicle users relies on manual labor to spend a lot of time collecting and analyzing, which reduces the efficiency of car insurance cost calculation.

Summary of the invention

According to various embodiments disclosed in the present application, a driving data analysis method, device, computer equipment, and storage medium are provided.

A driving data analysis method, executed by a computer device, the method comprising: acquiring driving data of a target vehicle; the driving data including a driving image; determining a recognition area in the driving image; identifying the presence of a vehicle identifier in the recognition area Nearby vehicles, record the vehicle position of the nearby vehicles; determine whether the target vehicle has overtaking behavior by comparing the changes in the vehicle positions of nearby vehicles in the adjacent multi-frame driving images; the overtaking frequency of the target vehicle according to the judgment result Performing statistics; and calculating the auto insurance cost corresponding to the target vehicle according to the overtaking frequency.

In one of the embodiments, the determining the recognition area in the driving image includes: recognizing the recognition starting point and the lane edge in the driving image; obtaining the following distance between the target vehicle and the preceding vehicle in the same lane, according to the The following distance determines the recognition distance; and the recognition area is determined based on the recognition starting point and the recognition distance.

In one of the embodiments, the recording the vehicle position of the nearby vehicles includes: generating a uniform edge line according to the identification distance; dividing the identification area into a plurality of sub-areas based on the uniform edge line and the lane edge; And determine the corresponding vehicle location according to the location of the nearby vehicle in the sub-region.

In one of the embodiments, determining whether the target vehicle has overtaking behavior by comparing changes in the vehicle position of nearby vehicles in adjacent multiple frames of driving images includes: generating according to the vehicle position in adjacent multiple frames of driving images The driving feature vector of nearby vehicles; calculating the first attribute value of the driving feature vector, and comparing whether the first attribute value reaches a threshold; if it reaches the threshold, calculating the second attribute value of the driving feature vector; judging the first attribute value Whether the second attribute value is a target attribute value; and if it is a target attribute value, marking the target vehicle has an overtaking behavior.

In one of the embodiments, the driving data includes driving time; generating driving feature vectors of nearby vehicles according to the position of the vehicle in the adjacent multiple frames of driving images includes: determining the driving time of the multi-frame driving image according to the driving time Traversal sequence; according to the traversal sequence, traverse each frame of traffic image in turn whether there are nearby vehicles; mark the vehicle position of the accessory vehicle in one or more frames of traffic image as vector elements in different sequences; for each accessory The adjacent vector elements of the vehicles are deduplicated; and based on the deduplicated multiple vector elements, the driving feature vectors of the corresponding nearby vehicles are generated.

In one of the embodiments, the driving data further includes vehicle sensing data; the calculating the car insurance cost corresponding to the target vehicle according to the overtaking frequency includes: identifying the lane departure frequency and collision of the target vehicle based on the driving image Early warning frequency; based on the vehicle induction data, count the speeding frequency and sharp turning frequency of the target vehicle; crawl the bad driving record of the target vehicle, and count the drunk driving frequency and responsible accident of the target vehicle based on the bad driving record Frequency; Determine the driving safety level of the target vehicle based on the frequency of overtaking, lane departure, collision warning, speeding, sharp turning, drunk driving, and responsible accident frequency during the statistical period; and according to the driving safety level Adjust the auto insurance cost of the target vehicle.

A driving data analysis device, the device comprising: a driving image processing module for acquiring driving data of a target vehicle; the driving data includes a driving image; identifying a recognition area in the driving image; identifying a vehicle identifier that appears in the recognition The nearby vehicles in the area record the vehicle position of the nearby vehicles; the overtaking behavior analysis module is used to judge whether the target vehicle has overtaking behavior by comparing the changes in the vehicle position of nearby vehicles in adjacent multiple frames of driving images; As a result, statistics are made on the overtaking frequency of the target vehicle; and an auto insurance cost calculation module for calculating the auto insurance cost corresponding to the target vehicle according to the overtaking frequency.

In one of the embodiments, the driving image processing module is also used to identify the recognition starting point and lane edges in the driving image; obtain the following distance between the target vehicle and the preceding vehicle in the same lane, and determine the recognition based on the following distance Distance; and the recognition area is determined based on the recognition starting point and the recognition distance.

A computer device includes a memory and one or more processors. The memory stores computer readable instructions. When the computer readable instructions are executed by the processor, the steps of the driving data analysis method provided in any embodiment of the present application are implemented.

One or more non-volatile computer-readable storage media storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors implement any one of the embodiments of the present application. Provide the steps of the driving data analysis method.

The details of one or more embodiments of the application are set forth in the following drawings and description. Other features and advantages of this application will become apparent from the description, drawings and claims.

Description of the drawings

In order to more clearly describe the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work.

Fig. 1 is an application scenario diagram of a driving data analysis method according to one or more embodiments.

Fig. 2 is a schematic flowchart of a method for analyzing driving data according to one or more embodiments.

Fig. 3A is a schematic diagram of a process of driving image processing according to one or more embodiments.

3B is a schematic diagram of another process of driving image processing according to one or more embodiments.

3C is a schematic diagram of another process of driving image processing according to one or more embodiments.

Fig. 4 is a schematic flowchart of the steps of determining overtaking behavior in one or more embodiments.

Fig. 5 is a structural block diagram of a driving data analysis device according to one or more embodiments.

Figure 6 is a block diagram of a computer device according to one or more embodiments.

Detailed ways

In order to make the technical solutions and advantages of the present application clearer, the following further describes the present application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the application, and not used to limit the application.

The driving data analysis method provided in this application can be applied to the application environment as shown in FIG. 1. The terminal 102 and the server 104 communicate through the network. The terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, and portable wearable devices. The terminal 102 can be a terminal corresponding to the target vehicle owner, or an insurance company that the target vehicle owner wants to handle auto insurance business. The corresponding terminal. The server 104 may be implemented as an independent server or a server cluster composed of multiple servers. When it is desired to handle auto insurance business based on the target vehicle, the user can send a request for auto insurance handling to the server based on the terminal 102. The vehicle insurance application request carries the target vehicle identification. The server 104 obtains driving data corresponding to the target vehicle according to the target vehicle identifier. The driving data includes multiple frames of driving images. The server 104 determines the recognition area in the driving image, recognizes the vehicle whose vehicle identifier appears in the recognition area, and marks the recognized vehicle as a nearby vehicle. The server 104 records the vehicle positions of nearby vehicles in adjacent multiple frames of driving images, and compares the changes of the vehicle positions, and can determine whether the target vehicle has overtaking behavior according to the comparison result. The server 104 counts the overtaking frequency of the target vehicle according to the judgment result, and can judge the safety of the driving behavior of the user of the target vehicle according to the overtaking frequency, and further calculates the auto insurance cost corresponding to the target vehicle according to the overtaking frequency. The server 104 returns the calculated car insurance cost to the terminal 102. The above-mentioned auto insurance cost calculation process automatically collects and analyzes the driving data, and directly adjusts the auto insurance cost according to the analysis result, which greatly reduces the labor burden. It not only improves the efficiency of auto insurance cost calculation, but also improves the objectivity and accuracy of auto insurance cost calculation.

In one of the embodiments, as shown in FIG. 2, a method for analyzing traffic data is provided. Taking the method applied to the server in FIG. 1 as an example, the method includes the following steps:

Step 202: Acquire driving data of the target vehicle; the driving data includes driving images.

The server makes full use of the driving data collected and recorded by the driving recorder, and collects driving data of the target vehicle according to the preset time frequency. The driving data includes multiple frames of driving images and the driving time corresponding to each frame of driving images.

Step 204: Determine the recognition area in the driving image.

The server determines a certain area around the target vehicle as the recognition area in the driving image. For example, an area with a preset area on at least one of the front, rear, left, or right side of the target vehicle may be used as the recognition area. The preset area may also be a fixed value or a dynamic value such as a preset ratio of the driving image.

Step 206: Recognize nearby vehicles whose vehicle identifiers appear in the recognition area, and record the vehicle positions of nearby vehicles.

The vehicle identification can be a license plate number or the like. The recognition area includes a plurality of sub-areas. The server determines the vehicle location of each nearby vehicle according to which sub-area of the recognition area the accessory vehicle is located.

Step 208: Determine whether the target vehicle has an overtaking behavior by comparing changes in the vehicle position of nearby vehicles in adjacent multiple frames of driving images.

The adjacent multi-frames may be the preset number of frames acquired recently, such as 3 frames. It is easy to understand that the number of frames of the comparison driving image can be set freely as needed, and there is no restriction on this. The license plate mark of some nearby vehicles may only appear in some frames of the driving image in the preset number of frames of the driving image. Each nearby vehicle is recognized for the first time with only one vehicle position. As the number of driving image frames increases, the stored vehicle positions gradually increase, but only a preset number of driving positions can be stored at most. In other words, the number of stored vehicle positions of each nearby vehicle is less than or equal to the preset number.

The server obtains the change trend of the vehicle position of each nearby vehicle in the multi-frame driving image, and determines whether the change trend is the preset first trend. If yes, the server determines that the target vehicle has overtaking behavior during the corresponding driving time.

Step 210: Count the overtaking frequency of the target vehicle according to the judgment result.

According to the judgment result, the server counts the overtaking frequency of the target vehicle during the statistical period. The statistical period may be a period of time before the owner of the target vehicle initiates the auto insurance application request, such as half a year. The frequency of overtaking can be the ratio of the number of overtakings to the length of the statistical period.

In another embodiment, the server counts the number of times the target vehicle has been overtaken during the statistical period (denoted as the number of overtakes). For example, the server determines whether the changing trend of the vehicle position of nearby vehicles in the multiple frames of driving images is a preset second trend. If yes, the server determines that the target vehicle has been overtaken during the corresponding driving time. At this time, the calculation of the overtaking frequency may be: overtaking frequency=number of overtaking/(number of overtaking+number of overtaking).

Step 212: Calculate the auto insurance cost corresponding to the target vehicle according to the frequency of overtaking.

The server can preset multiple overtaking frequency intervals and the auto insurance cost adjustment ratio corresponding to each overtaking frequency interval. The server determines the overtaking frequency interval to which the overtaking frequency of the target vehicle belongs, and increases or decreases the basic auto insurance expense according to the auto insurance cost adjustment ratio corresponding to the overtaking frequency interval to obtain the vehicle expense corresponding to the target vehicle.

In this embodiment, according to the acquired multiple frames of driving images of the target vehicle, the recognition area in the driving image can be determined; according to whether the license plate mark appears in the recognition area, the nearby vehicles corresponding to the target vehicle can be identified; The vehicle position in adjacent multiple frames of driving images can be compared to the changes in the vehicle position; according to the changes in the vehicle position, it can be judged whether the target vehicle has overtaking behavior; according to the judgment result, the overtaking frequency of the target vehicle can be calculated; according to the overtaking frequency, you can Calculate the auto insurance cost corresponding to the target vehicle. Since the collection and analysis of driving data are automatically performed, and the auto insurance costs are directly calculated based on the analysis results, not only the efficiency of calculating the auto insurance costs can be improved, but the objectivity and accuracy of the calculation results can also be improved. In addition, by dividing the recognition area of the driving image, and performing position statistics of nearby vehicles based on the recognition area, judging whether the target vehicle has overtaking behavior according to the position change of the nearby vehicle relative to the target vehicle, which can improve the similarity of the image compared to the general comparison. The accuracy of judging overtaking behaviors improves the accuracy of car insurance calculations.

In one of the embodiments, the step of determining the recognition area in the driving image includes: recognizing the recognition starting point and the lane edge in the driving image; obtaining the following distance between the target vehicle and the preceding vehicle in the same lane, and determining the recognition distance according to the following distance ; Determine the recognition area based on the recognition starting point and the recognition distance.

The driving recorder usually collects the vehicle conditions and road conditions around the target vehicle with the target vehicle as the center, so that the server can mark the lower position in the middle of the driving image (that is, the location of the target vehicle) as the identification starting point. There are usually lines, arrows, text, facade markings, raised road signs and outline signs on the road surface that are used to convey traffic information such as guidance, restrictions, and warnings to traffic participants. Among them, the lane edge refers to the line dividing the lane by the user on the road where the target vehicle is located.

If the target vehicle has a preceding vehicle in the same lane in the driving image, the server obtains the image distance of the target vehicle from the preceding vehicle in the same lane, and calculates the following distance of the target distance based on the image distance and the image shooting ratio. If the target vehicle does not exist in the same lane in the driving image, the server obtains the image shooting distance, and calculates the following distance of the target distance according to the image shooting distance and the image shooting ratio. The server performs a preset logical operation on the following distance to obtain the recognition distance. For example, following distance *3/2 = recognition distance. In another embodiment, the recognition distance can be dynamically determined by a preset ratio of the driving image, or it can be a fixed value, which is not limited.

The recognition area may be a quadrilateral whose side length is determined according to the preset length and the recognition distance. Among them, the identification starting point is the midpoint of one side of the quadrilateral. Fig. 3A is one frame of driving image taken by the driving recorder of the target vehicle. As shown in FIG. 3A, the recognition area corresponding to the driving image may be an isosceles trapezoid with the recognition starting point as the lower end point, wherein the lower side length and the upper side length may be the image lane width * 3, and the height may be the recognition distance. The image lane width may be the width of a lane in the driving image. It is easy to understand that in different image heights of the driving image, the corresponding image lane width is different. For example, the width of the image lane with the image height at the bottom side can be 5 cm; the width of the image lane with the image height at the top side can be 3 cm.

The server identifies nearby vehicles of the target vehicle in the driving image. In the above example, there are five vehicles in the recognition area. Among them, four vehicles A, B, C, and D can be recognized. Although vehicle E can recognize its license plate, it is not in the recognition area, so the target The nearby vehicles of the vehicle include A, B, C, and D.

In this embodiment, image processing is performed on the driving image to dynamically determine the recognition area of the target vehicle in each driving image. Compared with the general frame selection method, the accuracy of the area division can be improved; the recognition area can be dynamically limited to further Precisely limiting the content of detailed image processing not only improves the recognition accuracy, but also improves the efficiency of recognition because the amount of data that needs image processing is limited.

In one of the embodiments, recording the vehicle position of nearby vehicles includes: generating an even sideline based on the recognition distance; dividing the recognition area into multiple sub-areas based on the average sideline and lane sideline; and determining the correspondence based on the location of the sub-areas of nearby vehicles The location of the vehicle.

The number of equally divided edges varies according to the number of copies to be evenly divided into the recognition area. For example, the number of equally divided edges may be the number of copies to be divided into the recognition area -1. The length of the different equally divided sides can be different. The length of the even sideline can also be an integer multiple of the image lane width*3. For example, after performing area division on the driving image in the above example, an image as shown in FIG. 3B can be obtained. In FIG. 3B, the server divides the recognition area into three equal parts according to the recognition distance. Specifically, the server generates three equally-divided edges according to the recognition distance and the number of copies to be evenly divided into the recognition area. Wherein, the length of the divided sideline 1 may be the width of the image lane*3, and the lengths of the divided sideline 2 and the divided sideline 3 may be the width of the image lane*1, respectively.

The server may construct a coordinate system based on the recognition area, and then determine the image coordinates of the dividing edge and the lane edge respectively in the driving image. The server splices the multiple equally divided edges and lane edges in the recognition area according to the image coordinates, and then divides the recognition area into multiple sub-areas, and marks the multiple as upper area and the upper area according to the coordinate position of the sub-areas in the coordinate system. Middle area or lower area. As shown in Figure 3C, in the above example, the server divides the recognition area into 6 sub-areas, marks the sub-areas 1 with the largest ordinate as the upper area, and marks the sub-areas 2 and 3 with the second highest and the same ordinate as the upper area. In the middle area, the sub-area 5 and the sub-area 6 with the lowest ordinate and the same are marked as the lower area. It is easy to understand that if the number of copies of the recognition area is less than 3 copies, a certain sub-area can be further divided into multiple intermediate areas with different ordinates, and then divided into upper, middle, and lower according to the above method area. If the number of copies of the recognition area is more than 3 copies, multiple sub-areas adjacent to the ordinate can be merged into a middle area and then divided into the upper, middle and lower areas according to the above method.

According to the location of the nearby vehicle in the sub-area, the server can determine the corresponding vehicle location. For example, if the vehicle identifier of nearby vehicle A appears in the upper area, the vehicle position of nearby vehicles can be recorded as up. In another embodiment, the vehicle position of the nearby vehicle may be the specific coordinates of the coordinate system corresponding to the vehicle identifier of the nearby vehicle A in the recognition area, which is not limited.

It is worth noting that other methods can also be used to determine the recognition area, and other methods can also be used to divide the recognition area. This embodiment only provides an exemplary method that can determine the vehicle position change of nearby vehicles relative to the target vehicle.

In this embodiment, the recognition area is further divided into multiple sub-areas that facilitate the determination of the position changes of nearby vehicles relative to the target vehicle. Based on the multiple sub-areas divided in this way, the recognition algorithm for recognizing the change trend of nearby vehicles relative to the target vehicle position can be simplified. And then improve the efficiency of overtaking behavior analysis.

In one of the embodiments, as shown in FIG. 4, by comparing the changes in the vehicle positions of nearby vehicles in adjacent frames of driving images, it is determined whether the target vehicle has overtaking behavior, that is, the step of determining overtaking behavior includes:

Step 402: Generate driving feature vectors of nearby vehicles according to the vehicle positions in adjacent multiple frames of driving images.

In one of the embodiments, the driving data includes the driving time; according to the vehicle position in the adjacent multiple frames of driving images, generating the driving feature vector of nearby vehicles includes: determining the traversal sequence of the multiple frames of driving images according to the driving time; Traversal sequence, traverse each frame of the driving image in turn whether there are nearby vehicles; mark the vehicle position of the accessory vehicle in one or more frames of the driving image as vector elements in different orders; for the adjacent vector elements of each accessory vehicle Perform deduplication processing. Based on the multiple vector elements after deduplication, the driving feature vectors of the corresponding nearby vehicles are generated.

The server traverses the collected multiple frames of driving images in the order of driving time. It is easy to understand that some nearby vehicles may only appear in some frames of the traffic images in the recently collected multi-frame traffic images. Therefore, during the traversal process, the server judges whether the license plate identifiers of nearby vehicles exist in the first frame of traffic images collected. If it exists in the first frame of driving image, the server marks the vehicle position of the accessory vehicle in the first frame of driving image as a vector element in the first order. The server continues to determine whether the license plate identifiers of nearby vehicles are present in the next frame of traffic images collected. If it exists in the next frame of driving image, the server marks the vehicle position of the accessory vehicle in the next frame of driving image as the next sequential vector element, and continues to traverse to determine whether the license plate mark of nearby vehicles exists in the next frame of driving image. Repeat the traversal until the last frame of driving image, and obtain multiple vector elements corresponding to each nearby vehicle. If the next frame of traffic image does not exist, the server continues to traverse the next frame of traffic image in the above-mentioned manner.

According to the order of obtaining vector elements, the server sorts multiple vector elements to form an element queue. The server determines whether each vector element in the element queue is duplicated with the previous vector element. If repetition occurs, the server deletes the corresponding vector element from the element queue, and generates the driving feature vector of the corresponding nearby vehicle based on the deduplicated element queue. For example, in the above example, the driving feature vector corresponding to nearby vehicle A can be [up, middle, down], the driving feature vector corresponding to nearby vehicle B can be [down, middle, up], and the driving feature corresponding to nearby vehicle C The vector can be [up, middle], and the driving feature vector corresponding to the nearby vehicle D can be [middle, bottom]. It is easy to understand that there is no such as [middle, down, down] and similar traffic feature vectors with repeated adjacent vector elements.

Step 404: Calculate the first attribute value of the driving feature vector, and compare whether the first attribute value reaches the threshold.

The first attribute value may be the number of vector elements contained in the driving feature vector. The threshold can be a fixed value, such as 3. The threshold value may also be a value dynamically determined according to the number of copies of the driving image to be divided equally.

Step 406: If the threshold is reached, calculate the second attribute value of the driving feature vector.

If the first attribute value of the driving feature vector is less than the threshold, the server does not determine whether the target vehicle overtakes the corresponding nearby vehicle. For example, in the above example, the first attribute value of nearby vehicle C and nearby vehicle D is 2, which is less than the threshold value 3.

If the first attribute value of the driving feature vector is equal to the threshold, the server further calculates the second attribute value of the driving feature vector. The second attribute value may be an attribute value used to characterize the change trend of the vehicle position of nearby vehicles relative to the target vehicle.

Step 408: Determine whether the second attribute value is the target attribute value.

In step 410, if it is the target attribute value, mark that the target vehicle has an overtaking behavior.

The server presets multiple target attribute values and determination results associated with each target attribute value. For example, in the above example, the second attribute value corresponding to nearby vehicle A is the target attribute value 3, indicating that the target vehicle is moving forward relative to nearby vehicle A, and it can be determined that the target vehicle has overtaken relative to nearby vehicle A.

In this embodiment, the overtaking behavior analysis algorithm is simplified. By determining whether the first attribute value and the second attribute value of the driving feature vector meet the corresponding preset conditions respectively, it can be judged whether the target vehicle has overtaking behavior, and the overtaking behavior can be improved. Analyze efficiency.

In one of the embodiments, calculating the car insurance cost corresponding to the target vehicle based on the frequency of overtaking includes: recognizing the lane departure frequency and collision warning frequency of the target vehicle based on driving images; and counting the frequency of overspeeding and sharp turning of the target vehicle based on vehicle sensing data; Crawl the bad driving record of the target vehicle, and calculate the frequency of drunk driving and responsible accident frequency of the target vehicle based on the bad driving record; according to the frequency of overtaking, lane departure, collision warning frequency, speeding frequency, sharp turn frequency, drunk driving frequency and liability during the statistical period Accident frequency, determine the driving safety level of the target vehicle; adjust the auto insurance cost of the target vehicle according to the driving safety level.

The server determines whether the target vehicle has lane departure behavior by comparing the position changes of the target vehicle relative to the lane edge in the adjacent multiple frames of driving images. When there is lane departure behavior, the server counts the lane departure frequency of the target vehicle. The server determines whether the target vehicle has collision warning behavior by comparing whether the following distance is less than a preset value. When there is a collision warning behavior, the server counts the collision warning frequency of the target vehicle.

Driving data also includes vehicle sensing data. Vehicle sensing data includes speed change data and direction change data. The server obtains the corresponding speed limit data of the driving section according to the driving image, and determines whether the target vehicle has speeding behavior based on the speed limit data and the speed change data of the driving section. If there is speeding behavior, the server will make statistics on the speeding frequency of the target vehicle. The server determines whether the target vehicle has a sharp turn based on the direction change data. If there is a sharp turn, the server will make statistics on the frequency of sharp turns of the target vehicle.

The server crawls the poor driving record of the target vehicle on the traffic management website. Poor driving records include drunk driving records, and traffic accident records with full or partial responsibility. The server counts the frequency of drunk driving and the frequency of responsible accidents of the target vehicle based on the bad driving record.

The server corresponds to multiple dimensions of safety evaluation indicators based on the target vehicle's overtaking frequency, lane departure frequency, collision warning frequency, speeding frequency, sharp turning frequency, drunk driving frequency, and responsible accident frequency during the statistical period of the target vehicle, as well as the preset safety evaluation indicators of different dimensions The index weight of, can comprehensively determine the driving safety level of the target vehicle. According to the safety level of driving behavior, the auto insurance cost of the target vehicle can be adjusted. For example, setting corresponding different premium discount rates based on different overtaking frequency.

In this embodiment, based on multiple dimensions of safety evaluation indicators such as overtaking frequency, lane departure frequency, collision warning frequency, overspeed frequency, sharp turn frequency, drunk driving frequency, and responsible accident frequency, the safety level of driving behavior of the target vehicle is determined, which can improve driving The accuracy of the calculation of behavioral safety levels, thereby improving the accuracy of the calculation of auto insurance costs.

It should be understood that although the various steps in the flowcharts of FIGS. 2 and 4 are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless specifically stated in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least part of the steps in Figures 2 and 4 may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. These sub-steps or The order of execution of the stages is not necessarily carried out sequentially, but may be executed alternately or alternately with other steps or at least a part of the sub-steps or stages of other steps.

In one of the embodiments, as shown in FIG. 5, a driving data analysis device is provided, which includes a driving image processing module 502, an overtaking behavior analysis module 504, and a car insurance cost calculation module 506, wherein:

The driving image processing module 502 is used to obtain driving data of the target vehicle; the driving data includes driving images; identifying the recognition area in the driving image; identifying nearby vehicles with vehicle identifiers in the recognition area, and recording the vehicle positions of nearby vehicles.

The overtaking behavior analysis module 504 is used for judging whether the target vehicle has overtaking behavior by comparing changes in the vehicle position of nearby vehicles in the adjacent multiple frames of driving images; and counting the overtaking frequency of the target vehicle according to the judgment result.

The auto insurance cost calculation module 506 is used to calculate the auto insurance cost corresponding to the target vehicle according to the frequency of overtaking.

In one of the embodiments, the driving image processing module 502 is also used to identify the recognition starting point and the lane edge in the driving image; obtain the following distance between the target vehicle and the preceding vehicle in the same lane, and determine the recognition distance based on the following distance; And the recognition distance to determine the recognition area.

In one of the embodiments, the driving image processing module 502 is also used to generate an even sideline based on the recognition distance; divide the recognition area into multiple sub-areas based on the average sideline and lane sidelines; determine the corresponding sub-areas based on the location of the nearby vehicle Vehicle location.

In one of the embodiments, the overtaking behavior analysis module 504 is further configured to generate driving feature vectors of nearby vehicles according to the vehicle positions in the adjacent multiple frames of driving images; calculate the first attribute value of the driving feature vector, and compare the first attributes Whether the value reaches the threshold; if it reaches the threshold, calculate the second attribute value of the driving feature vector; determine whether the second attribute value is the target attribute value; if it is the target attribute value, mark the target vehicle as overtaking.

In one of the embodiments, the driving data includes driving time; the overtaking behavior analysis module 504 is also used to determine the traversal sequence of multiple frames of driving images according to the driving time; according to the traversal order, sequentially traverse whether each frame of driving image appears nearby vehicles ; Mark the vehicle position of the accessory vehicle in one or more frames of driving images as vector elements in different sequences; de-duplicate the adjacent vector elements of each accessory vehicle; generate based on multiple vector elements after de-duplication Corresponding to the driving feature vector of nearby vehicles.

In one of the embodiments, the driving data further includes vehicle sensing data; the car insurance cost calculation module 506 is also used for identifying the lane departure frequency and collision warning frequency of the target vehicle based on the driving image; and counting the speeding frequency and emergency of the target vehicle based on the vehicle sensing data. Turning frequency; crawl the bad driving record of the target vehicle, and calculate the frequency of drunk driving and responsible accident frequency of the target vehicle based on the bad driving record; according to the frequency of overtaking, lane departure, collision warning frequency, speeding frequency, sharp turning frequency, drunk driving during the statistical period The frequency and frequency of responsible accidents determine the driving safety level of the target vehicle; adjust the auto insurance cost of the target vehicle according to the driving safety level.

For the specific limitation of the driving data analysis device, please refer to the above limitation of the driving data analysis method, which will not be repeated here. Each module in the above-mentioned driving data analysis device can be implemented in whole or in part by software, hardware and a combination thereof. The foregoing modules may be embedded in the form of hardware or independent of the processor in the computer device, or may be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the foregoing modules.

In one of the embodiments, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. 6. The computer equipment includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor of the computer device is used to provide calculation and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer readable instructions, and a database. The internal memory provides an environment for the operation of the operating system and computer-readable instructions in the non-volatile storage medium. The database of the computer equipment is used to store the driving data of the target vehicle. The network interface of the computer device is used to communicate with an external terminal through a network connection. The computer readable instruction is executed by the processor to realize a driving data analysis method.

Those skilled in the art can understand that the structure shown in FIG. 6 is only a block diagram of part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may Including more or fewer parts than shown in the figure, or combining some parts, or having a different arrangement of parts.

One or more non-volatile storage media storing computer-readable instructions. When the computer-readable instructions are executed by one or more processors, the one or more processors implement the driving provided in any one of the embodiments of the present application. Steps of data analysis method.

Persons of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by computer-readable instructions to instruct relevant hardware. The computer-readable instructions can be stored in a non-volatile computer readable In the storage medium, when the computer-readable instructions are executed, they may include the procedures of the above-mentioned method embodiments. Wherein, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and/or volatile memory. Non-volatile memory may include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not a limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous chain Channel (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction between the combinations of these technical features, they should It is considered as the range described in this specification.

The above examples only express several implementation manners of this application, and the descriptions are more specific and detailed, but they cannot therefore be understood as limiting the scope of the invention patent. It should be pointed out that for those of ordinary skill in the art, without departing from the concept of this application, several modifications and improvements can be made, and these all fall within the protection scope of this application. Therefore, the scope of protection of the patent of this application shall be subject to the appended claims.

Claims

A method for analyzing driving data, which is executed by a computer device, and the method includes:

Acquiring driving data of the target vehicle; the driving data includes driving images;

Determining a recognition area in the driving image;

Identify nearby vehicles whose vehicle identifiers appear in the identification area, and record the vehicle positions of the nearby vehicles;

By comparing changes in the vehicle position of nearby vehicles in adjacent multiple frames of driving images, determine whether the target vehicle has overtaking behavior;

Count the overtaking frequency of the target vehicle according to the judgment result; and

Calculate the auto insurance cost corresponding to the target vehicle according to the overtaking frequency.
The method according to claim 1, wherein the determining the recognition area in the driving image comprises:

Recognizing the recognition starting point and lane edge in the driving image;

Obtain the following distance between the target vehicle and the preceding vehicle in the same lane, and determine the recognition distance according to the following distance; and

The recognition area is determined based on the recognition starting point and the recognition distance.
The method of claim 2, wherein the recording the vehicle position of the nearby vehicle comprises:

Generating an evenly divided side line according to the identification distance;

Dividing the recognition area into a plurality of sub-areas based on the equally divided sideline and the lane sideline; and

Determine the corresponding vehicle location based on the location of the nearby vehicle in the sub-region.
The method according to claim 1, wherein the judging whether the target vehicle has an overtaking behavior by comparing changes in the vehicle position of nearby vehicles in adjacent multiple frames of driving images comprises:

Generate the driving feature vector of nearby vehicles according to the vehicle position in the adjacent multiple frames of driving images;

Calculating the first attribute value of the driving feature vector, and comparing whether the first attribute value reaches a threshold;

If the threshold is reached, calculate the second attribute value of the driving feature vector;

Determine whether the second attribute value is the target attribute value; and

If it is the target attribute value, mark that the target vehicle has an overtaking behavior.
The method according to claim 4, wherein the driving data includes driving time; generating driving feature vectors of nearby vehicles according to the position of the vehicle in multiple adjacent frames of driving images, comprising:

Determining the traversal sequence of multiple frames of driving images according to the driving time;

According to the traversal sequence, sequentially traverse whether nearby vehicles appear in each frame of the driving image;

Mark the vehicle position of the accessory vehicle in one or more frames of driving images as vector elements in different sequences;

De-duplicate the adjacent vector elements of each accessory vehicle; and

Based on the multiple vector elements after deduplication, the driving feature vectors of the corresponding nearby vehicles are generated.
The method according to claim 1, wherein the driving data further includes vehicle sensing data; the calculating the auto insurance cost corresponding to the target vehicle according to the frequency of overtaking includes:

Identifying the lane departure frequency and collision warning frequency of the target vehicle based on the driving image;

Counting the frequency of overspeed and the frequency of sharp turns of the target vehicle based on the vehicle induction data;

Crawling the bad driving record of the target vehicle, and counting the frequency of drunk driving and the frequency of responsible accidents of the target vehicle based on the bad driving record;

Determine the driving safety level of the target vehicle according to the frequency of overtaking, frequency of lane departure, frequency of collision warning, frequency of overspeed, frequency of sharp turns, frequency of drunk driving, and frequency of responsible accidents during the statistical period; and

Adjust the auto insurance cost of the target vehicle according to the safety level of the driving behavior.
A driving data analysis device, the device includes:

The driving image processing module is used to obtain driving data of the target vehicle; the driving data includes driving images; identifying the recognition area in the driving image; identifying nearby vehicles whose vehicle identifiers appear in the recognition area, and recording the vehicles of the nearby vehicles position;

The overtaking behavior analysis module is used for judging whether the target vehicle has overtaking behavior by comparing changes in the vehicle position of nearby vehicles in adjacent multiple frames of driving images; and calculating the overtaking frequency of the target vehicle according to the judgment result;

The auto insurance cost calculation module is used to calculate the auto insurance cost corresponding to the target vehicle according to the overtaking frequency.
The device according to claim 7, wherein the driving image processing module is further used to identify the recognition starting point and the lane edge in the driving image; to obtain the following distance between the target vehicle and the preceding vehicle in the same lane, according to the The following distance determines the recognition distance; the recognition area is determined based on the recognition starting point and the recognition distance.
A computer device includes a memory and one or more processors. The memory stores computer-readable instructions. When the computer-readable instructions are executed by the one or more processors, the one or more Each processor performs the following steps:

Acquiring driving data of the target vehicle; the driving data includes driving images;

Determining a recognition area in the driving image;

Identify nearby vehicles whose vehicle identifiers appear in the identification area, and record the vehicle positions of the nearby vehicles;

By comparing changes in the vehicle position of nearby vehicles in adjacent multiple frames of driving images, determine whether the target vehicle has overtaking behavior;

Count the overtaking frequency of the target vehicle according to the judgment result; and

Calculate the auto insurance cost corresponding to the target vehicle according to the overtaking frequency.
The computer device according to claim 9, wherein the processor further executes the following steps when executing the computer-readable instruction:

Recognizing the recognition starting point and lane edge in the driving image;

Obtain the following distance between the target vehicle and the preceding vehicle in the same lane, and determine the recognition distance according to the following distance; and

The recognition area is determined based on the recognition starting point and the recognition distance.
The computer device according to claim 10, wherein the processor further executes the following steps when executing the computer-readable instruction:

Generating an evenly divided side line according to the identification distance;

Dividing the recognition area into a plurality of sub-areas based on the equally divided sideline and the lane sideline; and

Determine the corresponding vehicle location based on the location of the nearby vehicle in the sub-region.
The computer device according to claim 9, wherein the processor further executes the following steps when executing the computer-readable instruction:

Generate the driving feature vector of nearby vehicles according to the vehicle position in the adjacent multiple frames of driving images;

Calculating the first attribute value of the driving feature vector, and comparing whether the first attribute value reaches a threshold;

If the threshold is reached, calculate the second attribute value of the driving feature vector;

Determine whether the second attribute value is the target attribute value; and

If it is the target attribute value, mark that the target vehicle has an overtaking behavior.
The computer device according to claim 12, wherein the travel data includes travel time; the processor further executes the following steps when executing the computer readable instruction:

Determining the traversal sequence of multiple frames of driving images according to the driving time;

According to the traversal sequence, sequentially traverse whether nearby vehicles appear in each frame of the driving image;

Mark the vehicle position of the accessory vehicle in one or more frames of driving images as vector elements in different sequences;

De-duplicate the adjacent vector elements of each accessory vehicle; and

Based on the multiple vector elements after deduplication, the driving feature vectors of the corresponding nearby vehicles are generated.
The computer device according to claim 9, wherein the driving data further includes vehicle sensing data; the processor further executes the following steps when executing the computer-readable instruction:

Identifying the lane departure frequency and collision warning frequency of the target vehicle based on the driving image;

Counting the frequency of overspeed and the frequency of sharp turns of the target vehicle based on the vehicle induction data;

Crawling the bad driving record of the target vehicle, and counting the frequency of drunk driving and the frequency of responsible accidents of the target vehicle based on the bad driving record;

Determine the driving safety level of the target vehicle according to the frequency of overtaking, frequency of lane departure, frequency of collision warning, frequency of overspeed, frequency of sharp turns, frequency of drunk driving, and frequency of responsible accidents during the statistical period; and

Adjust the auto insurance cost of the target vehicle according to the safety level of the driving behavior.
One or more non-volatile computer-readable storage media storing computer-readable instructions, which when executed by one or more processors, cause the one or more processors to perform the following steps:

Acquiring driving data of the target vehicle; the driving data includes driving images;

Determining a recognition area in the driving image;

Identify nearby vehicles whose vehicle identifiers appear in the identification area, and record the vehicle positions of the nearby vehicles;

By comparing changes in the vehicle position of nearby vehicles in adjacent multiple frames of driving images, determine whether the target vehicle has overtaking behavior;

Count the overtaking frequency of the target vehicle according to the judgment result; and

Calculate the auto insurance cost corresponding to the target vehicle according to the overtaking frequency.
The storage medium according to claim 15, wherein the following steps are further executed when the computer-readable instructions are executed by the processor:

Recognizing the recognition starting point and lane edge in the driving image;

Obtain the following distance between the target vehicle and the preceding vehicle in the same lane, and determine the recognition distance according to the following distance; and

The recognition area is determined based on the recognition starting point and the recognition distance.
The storage medium according to claim 16, wherein the following steps are further executed when the computer-readable instructions are executed by the processor:

Generating an evenly divided side line according to the identification distance;

Dividing the recognition area into a plurality of sub-areas based on the equally divided sideline and the lane sideline; and

Determine the corresponding vehicle location based on the location of the nearby vehicle in the sub-region.
The storage medium according to claim 15, wherein the following steps are further executed when the computer-readable instructions are executed by the processor:

Generate the driving feature vector of nearby vehicles according to the vehicle position in the adjacent multiple frames of driving images;

Calculating the first attribute value of the driving feature vector, and comparing whether the first attribute value reaches a threshold;

If the threshold is reached, calculate the second attribute value of the driving feature vector;

Determine whether the second attribute value is the target attribute value; and

If it is the target attribute value, mark that the target vehicle has an overtaking behavior.
18. The storage medium according to claim 18, wherein the travel data includes travel time; and the following steps are further performed when the computer-readable instructions are executed by the processor:

Determining the traversal sequence of multiple frames of driving images according to the driving time;

According to the traversal sequence, sequentially traverse whether nearby vehicles appear in each frame of the driving image;

Mark the vehicle position of the accessory vehicle in one or more frames of driving images as vector elements in different sequences;

De-duplicate the adjacent vector elements of each accessory vehicle; and

Based on the multiple vector elements after deduplication, the driving feature vectors of the corresponding nearby vehicles are generated.
15. The storage medium according to claim 15, wherein the driving data further comprises vehicle sensing data; and the following steps are further executed when the computer-readable instructions are executed by the processor:

Identifying the lane departure frequency and collision warning frequency of the target vehicle based on the driving image;

Counting the frequency of overspeed and the frequency of sharp turns of the target vehicle based on the vehicle induction data;

Crawling the bad driving record of the target vehicle, and counting the frequency of drunk driving and the frequency of responsible accidents of the target vehicle based on the bad driving record;

Determine the driving safety level of the target vehicle according to the frequency of overtaking, frequency of lane departure, frequency of collision warning, frequency of overspeed, frequency of sharp turns, frequency of drunk driving, and frequency of responsible accidents during the statistical period; and

Adjust the auto insurance cost of the target vehicle according to the safety level of the driving behavior.