Sample acquisition method of driving behavior analysis model and related product

Abstract

The invention discloses a sample acquisition method for a driving behavior analysis model and a related product, and the method comprises the steps: determining a detection main body and a target conversion vehicle in a first sample, and enabling the target conversion vehicle to be a first environment vehicle adjacent to the detection main body; acquiring first pose information of the detection main body and second pose information of a target conversion vehicle; obtaining driving behavior data of the detection main body and driving behavior data of a target conversion vehicle; performing pose conversion on the detection main body and the target conversion vehicle according to the first pose information and the second pose information; and converting the driving behavior data of the detection main body and the target conversion vehicle to obtain a second sample taking the target conversion vehicle as a new detection main body. A large number of second samples are mined from existing first samples, and the cost of obtaining a large number of driving behavior analysis model samples is reduced.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a sample acquisition method for a driving behavior analysis model and related products. Background Art

[0002] As the level of autonomous driving functions increases, the requirements for simulation testing also increase. For traditional scenario simulation testing, the vehicle behavior in the simulation environment is fixed and the trajectory is simple, which is far from the actual driving environment. In order to improve the realism of the simulation test environment, it is necessary to collect driving behavior data in different scenarios from the actual driving environment to train the driving behavior analysis model. The more samples the training model has, the more realistic the fitted driver's driving behavior will be.

[0003] At present, the training of driving behavior analysis models is based on the driving behavior data of the detection subject as samples. It is necessary to collect driving behavior data under different driving environments from the detection subject to form different samples. However, the cost of obtaining samples for driving behavior analysis models through real vehicle data collection is high. Summary of the invention

[0004] Based on the above problems, the present application provides a sample acquisition method and related products of a driving behavior analysis model to reduce the cost of acquiring samples of the driving behavior analysis model.

[0005] The embodiments of the present application disclose the following technical solutions:

[0006] In a first aspect, an embodiment of the present application provides a method for acquiring samples of a driving behavior analysis model, the method comprising:

[0007] Determine a detection subject and a target conversion vehicle in a first sample, wherein the target conversion vehicle is a first environment vehicle adjacent to the detection subject;

[0008] Acquire the first posture information of the detection subject and the second posture information of the target conversion vehicle;

[0009] Acquiring driving behavior data of the detection subject and driving behavior data of the target conversion vehicle;

[0010] The detection subject and the target conversion vehicle are converted in position according to the first position information and the second position information; the driving behavior data of the detection subject and the target conversion vehicle are converted to obtain a second sample with the target conversion vehicle as a new detection subject.

[0011] Optionally, performing posture conversion on the detection subject and the target conversion vehicle according to the first posture information and the second posture information specifically includes:

[0012] The first coordinate system of the detection subject is converted into the coordinate system where the target conversion vehicle is located to obtain a second coordinate system;

[0013] In the second coordinate system, the fifth posture information of the target conversion vehicle as a new detection subject is obtained according to the coordinate conversion formula and the second posture information; the fourth posture information of the detection subject as an environment vehicle is obtained according to the coordinate conversion formula and the first posture information.

[0014] Optionally, the acquiring the driving behavior data of the detection subject and the driving behavior data of the target conversion vehicle specifically includes:

[0015] Acquire a vehicle driving state of the detection subject, where the vehicle driving state is one of the driving behavior data;

[0016] Determine a lane line between the detection subject and the target conversion vehicle that is closest to the target conversion vehicle according to the second posture information, and determine a lane line equation corresponding to the lane line;

[0017] Obtaining a distance between the target conversion vehicle and the lane line according to the second posture information and the lane line equation;

[0018] The vehicle driving state of the target conversion vehicle is obtained according to the distance.

[0019] Optionally, the first sample also includes a second environment vehicle adjacent to the detection subject, and the posture information of the second environment vehicle is third posture information. The method further includes:

[0020] In the second coordinate system, sixth posture information of the second environment vehicle is obtained according to the coordinate conversion formula and the third posture information.

[0021] Optionally, after obtaining the sample with the target conversion vehicle as the new detection subject, the method further includes:

[0022] Acquire a first duration that the target conversion vehicle is adjacent to the detection subject;

[0023] If the first duration is less than a preset threshold, the target conversion vehicle is deleted as a second sample of a new detection subject.

[0024] Optionally, the method further comprises:

[0025] Acquire a second duration that the second environment vehicle is adjacent to the detection subject;

[0026] If the second duration is less than a preset threshold, the second environment vehicle is deleted from the second sample.

[0027] Optionally, the driving behavior data includes speed and acceleration.

[0028] In a second aspect, an embodiment of the present application provides a sample acquisition device for a driving behavior analysis model, the device comprising: a determination module, a first acquisition module, a second acquisition module and a conversion module;

[0029] The determination module is used to determine a detection subject and a target conversion vehicle in a first sample, wherein the target conversion vehicle is a first environment vehicle adjacent to the detection subject;

[0030] The first acquisition module is used to acquire the first posture information of the detection subject and the second posture information of the target conversion vehicle,

[0031] The second acquisition module is used to acquire the driving behavior data of the detection subject and the driving behavior data of the target conversion vehicle;

[0032] The conversion module is used to perform posture conversion on the detection subject and the target conversion vehicle according to the first posture information and the second posture information; and convert the driving behavior data of the detection subject and the target conversion vehicle to obtain a second sample with the target conversion vehicle as the new detection subject.

[0033] Optionally, the conversion module includes a first conversion unit and a second conversion unit;

[0034] The first conversion unit is used to convert the first coordinate system of the detection subject into the coordinate system where the target conversion vehicle is located to obtain a second coordinate system;

[0035] In the second coordinate system, the fifth posture information of the target conversion vehicle as a new detection subject is obtained according to the coordinate conversion formula and the second posture information; the fourth posture information of the detection subject as an environment vehicle is obtained according to the coordinate conversion formula and the first posture information.

[0036] Optionally, the second acquisition module is specifically used to:

[0037] Acquire a vehicle driving state of the detection subject, where the vehicle driving state is one of the driving behavior data;

[0038] Determine a lane line between the detection subject and the target conversion vehicle that is closest to the target conversion vehicle according to the second posture information, and determine a lane line equation corresponding to the lane line;

[0039] Obtaining a distance between the target conversion vehicle and the lane line according to the second posture information and the lane line equation;

[0040] The vehicle driving state of the target conversion vehicle is obtained according to the distance.

[0041] Optionally, the first conversion unit is further used for:

[0042] In the second coordinate system, sixth posture information of the second environment vehicle is obtained according to the coordinate conversion formula and the third posture information.

[0043] Optionally, the device further comprises: a first deletion module;

[0044] The first deletion module is used to obtain a first duration of the target conversion vehicle being adjacent to the detection subject; if the first duration is less than a preset threshold, the target conversion vehicle is deleted as a second sample of a new detection subject.

[0045] Optionally, the device further comprises: a second deletion module;

[0046] The second deletion module is used to obtain a second duration that the second environment vehicle is adjacent to the detection subject; if the second duration is less than a preset threshold, the second environment vehicle is deleted from the second sample.

[0047] Optionally, the driving behavior data includes speed and acceleration.

[0048] In a third aspect, an embodiment of the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the sample acquisition method for the driving behavior analysis model as described in any one of the first aspects is implemented.

[0049] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which instructions are stored. When the instructions are executed on a terminal device, the terminal device executes the sample acquisition method of the driving behavior analysis model as described in any one of the first aspects.

[0050] Compared with the prior art, this application has the following beneficial effects:

[0051] First, determine the detection subject and the target conversion vehicle in the first sample, where the target conversion vehicle is the first environmental vehicle adjacent to the detection subject; secondly, obtain the first posture information of the detection subject and the second posture information of the target conversion vehicle; then, obtain the driving behavior data of the detection subject and the driving behavior data of the target conversion vehicle; finally, perform posture conversion on the detection subject and the target conversion vehicle according to the first posture information and the second posture information; convert the driving behavior data of the detection subject and the target conversion vehicle to obtain a second sample with the target conversion vehicle as the new detection subject. The present application obtains a number of samples equal to the number of environmental vehicles by performing position conversion and driving behavior data conversion on the detection subject and the environmental vehicles. Since the present application mines a large number of samples from existing samples, the cost of obtaining a large number of driving behavior analysis model samples is reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0053] Figure 1 A flow chart of a sample acquisition method for a driving behavior analysis model provided in an embodiment of the present application;

[0054] Figure 2 A schematic diagram of determining a target conversion vehicle provided in an embodiment of the present application;

[0055] Figure 3 A schematic diagram of a vehicle posture conversion provided in an embodiment of the present application;

[0056] Figure 4 A schematic diagram of obtaining a driving behavior analysis model sample provided in an embodiment of the present application;

[0057] Figure 5 A flowchart of another method for acquiring samples of a driving behavior analysis model provided in an embodiment of the present application;

[0058] Figure 6 A schematic diagram of determining a lane line equation provided in an embodiment of the present application;

[0059] Figure 7 A schematic diagram of obtaining the distance between a vehicle and a lane line provided in an embodiment of the present application;

[0060] Figure 8 A schematic diagram of the structure of a sample acquisition device for a driving behavior analysis model provided in an embodiment of the present application. DETAILED DESCRIPTION

[0061] As described above, currently, the driving behavior data of drivers in different scenarios is collected in real driving environments to train models that can characterize driver behavior. Since training models requires a large number of samples, if all samples are collected in real driving environments, the cost of training models will be very high.

[0062] To this end, the embodiment of the present application provides a sample acquisition method for a driving behavior analysis model. First, the detection subject and the target conversion vehicle in the first sample are determined, and the target conversion vehicle is the first environmental vehicle adjacent to the detection subject; secondly, the first posture information of the detection subject and the second posture information of the target conversion vehicle are obtained; then, the driving behavior data of the detection subject and the driving behavior data of the target conversion vehicle are obtained; finally, the detection subject and the target conversion vehicle are converted according to the first posture information and the second posture information; the driving behavior data of the detection subject and the target conversion vehicle are converted to obtain a second sample with the target conversion vehicle as the new detection subject.

[0063] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0064] See also Figure 1 , which is a flow chart of a sample acquisition method for a driving behavior analysis model provided in an embodiment of the present application.

[0065] like Figure 1 As shown, the method includes:

[0066] S101: Determine a detection subject and a target conversion vehicle in a first sample, where the target conversion vehicle is a first environment vehicle adjacent to the detection subject.

[0067] like Figure 2 As shown, in Figure 2 The medium-dark gray vehicle is the detection subject, and the remaining vehicles are all environmental vehicles. Any environmental vehicle within the dotted line can be used as a target conversion vehicle, and the environmental vehicle used as a target conversion vehicle is called the first environmental vehicle.

[0068] It can be understood that “adjacent” in the embodiment of the present application may refer to no obstructing objects between the two vehicles.

[0069] S102: Acquire the first posture information of the detection subject and the second posture information of the target conversion vehicle.

[0070] The first pose information may be the lateral distance, longitudinal distance and heading angle of the detection subject in the detection subject coordinate system. The second pose information may be the lateral distance, longitudinal distance and heading angle of the target conversion vehicle compared to the detection subject. The second pose information is a relative quantity.

[0071] It can be understood that the first posture information and the second posture information are both obtained by reading the memory or database inside the detection body. The difference is that the first posture information is the data of the detection body itself, and the second posture information is the data about the target conversion vehicle. The collection method can be through various sensors installed on the detection body, or it can be collected and obtained through a satellite system.

[0072] As an example, Figure 3 The lateral distance x1, longitudinal distance y1 and heading angle θ1 in the first pose information of the detected subject, that is, vehicle 1, are all 0. The first pose information can be expressed as:

[0073]

[0074] Figure 3 The second position information of the target conversion vehicle, that is, vehicle 3, can be expressed as:

[0075]

[0076] S103: Acquire the driving behavior data of the detection subject and the driving behavior data of the target conversion vehicle.

[0077] Among them, the driving behavior data of the detection subject is obtained based on the detection subject itself, while the driving behavior data of the target conversion vehicle is the relevant data of the target conversion vehicle collected by the detection subject.

[0078] As an example, when the driving behavior data is speed and acceleration, the speed and acceleration of the detection subject can be directly obtained; while the speed and acceleration of the target conversion vehicle are obtained through the sensor installed on the detection subject.

[0079] S104: Performing posture conversion on the detection subject and the target conversion vehicle according to the first posture information and the second posture information; converting the driving behavior data of the detection subject and the target conversion vehicle to obtain a second sample with the target conversion vehicle as a new detection subject.

[0080] S104 will be explained in the following three steps:

[0081] Step 1: Convert the first coordinate system of the detection subject to the coordinate system of the target conversion vehicle to obtain the second coordinate system.

[0082] The first coordinate system of the detection subject is Figure 3 As shown in (a) in the figure, the second coordinate system is Figure 3 As shown in (b) in .

[0083] The coordinate transformation formula is as follows:

[0084]

[0085] In the above conversion formula, t x ,t y The values ​​of are the coordinates of the origin of the original coordinate system in the new coordinate system, t θ is the angle between the second coordinate system and the first coordinate system; x, y, θ are the origin of the first coordinate system and the vehicle orientation.

[0086] Step 2: In the second coordinate system, the fifth posture information of the target conversion vehicle as the new detection subject is obtained according to the coordinate conversion formula and the second posture information; the fourth posture information of the detection subject as the environment vehicle is obtained according to the coordinate conversion formula and the first posture information.

[0087] When converting the first coordinate system to the second coordinate system, t in the transformation matrix θ =θ 13 , t x =x 13 , t y =y 13 , x, y, θ are the posture information of the detection subject.

[0088] The pose information of the target conversion vehicle after position conversion, that is, the fifth pose information is:

[0089]

[0090] The posture information of the original detection subject after position conversion, that is, the fourth posture information is:

[0091]

[0092] The above operations realize the position conversion between the detection subject and the target conversion vehicle.

[0093] Step 3: Convert the driving behavior data of the detection subject and the target vehicle.

[0094] It can be understood that the number of samples obtained by converting the vehicle position and the driving behavior data is equal to the number of adjacent environmental vehicles. That is, when the number of adjacent environmental vehicles is 4, 4 new samples can be obtained through S101-S104, such as Figure 4 That is, by using existing samples to mine a large number of samples, that is, the detection subject and the environment vehicle are converted in position and driving behavior data, and the number of samples equal to the number of environment vehicles is obtained. Since the application mines a large number of samples from existing samples, the cost of obtaining a large number of samples is reduced.

[0095] See also Figure 5 , which is a flowchart of another sample acquisition method for a driving behavior analysis model provided in an embodiment of the present application.

[0096] like Figure 5 As shown, the method includes:

[0097] S501: Determine the detection subject, target conversion vehicle and other environmental vehicles in the first sample.

[0098] The target conversion vehicle is a first environment vehicle adjacent to the detection subject, and the environment vehicles other than the first environment vehicle are regarded as other environment vehicles, namely, second environment vehicles.

[0099] S502: Acquire the first posture information of the detection subject, the second posture information of the target conversion vehicle, and the third posture information of the second environment vehicle.

[0100] It is understandable that the third posture information, like the second posture information, is data collected by the detection subject, wherein the collection method may be through various sensors installed on the detection subject, or may be collected and obtained through a satellite system.

[0101] As an example, the third pose information, i.e. Figure 3 The position information of vehicle 2 in can be expressed as:

[0102]

[0103] S503: Acquire the driving behavior data of the detection subject, the driving behavior data of the target conversion vehicle, and the driving behavior data of the second environment vehicle.

[0104] The method includes obtaining the vehicle driving state of the detection subject, the vehicle driving state of the target conversion vehicle and the vehicle driving state of the second environment vehicle. The vehicle driving state is one of the driving behavior data.

[0105] The vehicle driving status of the detection body can be obtained by directly reading the memory or database inside the detection body.

[0106] A method for obtaining the driving status of the target conversion vehicle and the second environment vehicle is described below using the target conversion vehicle as an example and divided into the following three steps:

[0107] Step 1: Determine the lane line closest to the target conversion vehicle between the detection subject and the target conversion vehicle according to the second posture information, and determine the lane line equation corresponding to the lane line.

[0108] The lane line closest to the target conversion vehicle between the detection subject and the target conversion vehicle is determined according to the lateral distance in the second posture information, and the corresponding lane line equation is determined.

[0109] The lane line equation can be expressed by the following formula:

[0110] y=C0+C1*x+C2*x 2 +C3*x 3

[0111] In the above formula, C0 is the offset distance of the vehicle from the lane boundary, C1 represents the lane line yaw angle, C2 represents the lane line curvature and is positive when it bends to the right, and C3 represents the lane line curvature change rate and when it is positive, it means that the curvature radius gradually decreases, and when it is negative, it means that the curvature radius gradually increases. The relationship between the lane line and x and y in the above formula is as follows: Figure 6 shown.

[0112] As an example, in the embodiment of the present application, the distance between the target conversion vehicle and the lane line is taken as an example, and the lateral distance between the target conversion vehicle and the detection subject is used to infer which lane the target conversion vehicle is located in, and then the lane line equation closest to the detection subject and the environment vehicle 2 is selected. Figure 7 As shown, in Figure 7 In the medium environment, vehicle 2 is located in the first lane to the right of the main vehicle. The lane line equation of lane line 2 is y2=C0+C1*x+C2*x. 2 +C3*x 3 .

[0113] Step 2: Obtain the distance between the target conversion vehicle and the lane line based on the second posture information and the lane line equation.

[0114] For ease of understanding, in the embodiment of the present application, in order to obtain vehicle 3 (i.e. Figure 3 Take the distance between vehicle 3) and the lane line as an example. The second posture information of vehicle 3 is as follows:

[0115]

[0116] x 13 Substituting into the cubic polynomial equation of lane line 2, we get the following formula:

[0117] y2=C0+C1*x 13 +C2*x 13 2 +C3*x 13 3

[0118] Through the above formula, we can get Figure 7 The value of d1 in Figure 7 D in is approximately equal to d2, so we can use d2=|y 13 The distance between the environment vehicle 2 and the lane line 2 is represented by |-|d1|.

[0119] Step 3: Obtain the vehicle driving state of the target conversion vehicle according to the distance.

[0120] Specifically, since the width of the lane line is a standard value, the distance between the environment vehicle 2 and another lane line 3 in the lane where the vehicle is located can be directly obtained by subtracting the value of d2 from the width of the standard lane.

[0121] In an optional implementation, the widths of lane 2 and lane 3 are calculated by subtracting the C0 coefficients in the lane line equations of lane 2 and lane 3.

[0122] By obtaining the distance between the environment vehicle (including the target conversion vehicle) and the lane line, it is possible to determine whether the vehicle is patrolling or changing lanes, thereby obtaining more accurate driving behavior data, that is, when the distance between the environment vehicle and the lane line is greater than the preset threshold, the environment vehicle may patrol the lane; when the distance between the environment vehicle and the lane line is less than the preset threshold, the environment vehicle may change lanes. When this sample is used in training the driving behavior analysis model, the realism of the simulation test environment can be improved.

[0123] S504: Perform posture conversion on the detection subject and the target conversion vehicle according to the first posture information and the second posture information; convert the driving behavior data of the detection subject and the target conversion vehicle to obtain a second sample with the target conversion vehicle as the new detection subject.

[0124] Step 1: Convert the first coordinate system of the detection subject to the coordinate system of the target conversion vehicle to obtain the second coordinate system.

[0125] Step 2: In the second coordinate system, according to the coordinate conversion formula and the second posture information, obtain the fifth posture information of the target conversion vehicle as the new detection subject; according to the coordinate conversion formula and the first posture information, obtain the fourth posture information of the detection subject as the environment vehicle; according to the coordinate conversion formula and the third posture information, obtain the sixth posture information of the second environment vehicle.

[0126] Among them, the second environment vehicle, namely Figure 3 The sixth posture information of vehicle 2 in is:

[0127]

[0128] S505: Obtain a first duration of time that the target conversion vehicle is adjacent to the detection subject; if the first duration is less than a preset threshold, delete the target conversion vehicle as a second sample of the new detection subject.

[0129] For the target conversion vehicles with a short life cycle, that is, the first duration, the amount of data characterizing their behavior is small and cannot be used as a complete sample, so the second sample with a short life cycle must be screened.

[0130] As an example, after executing S501-S505, four second samples are derived from the first sample, and the first time when the target conversion vehicle is adjacent to the detection subject is obtained in these four samples. If there are only three second samples whose first time is greater than the preset threshold, these three second samples are used for training the driving behavior analysis model, and the other sample is deleted.

[0131] S506: Obtain a second duration that the second environment vehicle is adjacent to the detection subject; if the second duration is less than a preset threshold, delete the second environment vehicle from the second sample.

[0132] Since the information of the second environment vehicle is collected by the main vehicle, the second environment vehicle that is active at the detection edge of the main vehicle will be considered as a new target object by the main vehicle each time it re-enters the detection range. The environment vehicle that is blocked during the detection process will also be considered as a new target object when it is detected again. Therefore, the ID of the same target will jump, resulting in many IDs with short life cycles in the sample. Therefore, these second environment vehicles with short cycles need to be filtered out after coordinate conversion.

[0133] As an example, in one of the second samples after S505, there are four environmental vehicles adjacent to the detection subject, and the environmental vehicles whose second duration of being adjacent to the detection subject is less than a preset threshold are deleted from the second sample.

[0134] It is understandable that this step will not reduce the number of second samples, but will screen the environmental vehicles in the samples, thereby improving the quality of the second samples and improving the accuracy of driving behavior analysis model training.

[0135] See also Figure 8 , Figure 8 A sample acquisition device for a driving behavior analysis model is provided for an embodiment of the present application, the device comprising: a determination module 801, a first acquisition module 802, a second acquisition module 803 and a conversion module 804;

[0136] A determination module 801 is used to determine a detection subject and a target conversion vehicle in a first sample, where the target conversion vehicle is a first environment vehicle adjacent to the detection subject;

[0137] The first acquisition module 802 is used to acquire the first posture information of the detection subject and the second posture information of the target conversion vehicle.

[0138] The second acquisition module 803 is used to acquire the driving behavior data of the detection subject and the driving behavior data of the target conversion vehicle;

[0139] The conversion module 804 is used to convert the posture of the detection subject and the target conversion vehicle according to the first posture information and the second posture information; convert the driving behavior data of the detection subject and the target conversion vehicle to obtain a second sample with the target conversion vehicle as the new detection subject.

[0140] Optionally, the conversion module 804 includes a first conversion unit and a second conversion unit;

[0141] A first conversion unit, used for converting the first coordinate system of the detection subject into the coordinate system where the target conversion vehicle is located to obtain a second coordinate system;

[0142] In the second coordinate system, the fifth posture information of the target conversion vehicle as the new detection subject is obtained according to the coordinate conversion formula and the second posture information; the fourth posture information of the detection subject as the environment vehicle is obtained according to the coordinate conversion formula and the first posture information.

[0143] Optionally, the second acquisition module 803 is specifically configured to:

[0144] Acquire the vehicle driving state of the detection subject, where the vehicle driving state is one type of driving behavior data;

[0145] Determine a lane line between the detection subject and the target conversion vehicle that is closest to the target conversion vehicle according to the second posture information, and determine a lane line equation corresponding to the lane line;

[0146] The distance between the target conversion vehicle and the lane line is obtained according to the second posture information and the lane line equation;

[0147] The vehicle driving state of the target conversion vehicle is obtained according to the distance.

[0148] Optionally, the first conversion unit is further configured to:

[0149] In the second coordinate system, the sixth posture information of the vehicle in the second environment is obtained according to the coordinate conversion formula and the third posture information.

[0150] Optionally, the device further includes: a first deletion module;

[0151] The first deletion module is used to obtain a first duration of time when the target conversion vehicle is adjacent to the detection subject; if the first duration is less than a preset threshold, the target conversion vehicle is deleted as a second sample of the new detection subject.

[0152] Optionally, the device further comprises: a second deletion module;

[0153] The second deletion module is used to obtain a second duration that the second environment vehicle is adjacent to the detection subject; if the second duration is less than a preset threshold, the second environment vehicle is deleted from the second sample.

[0154] Optionally, the driving behavior data includes speed and acceleration.

[0155] In addition, embodiments of the present application provide a computer device and a computer-readable storage medium.

[0156] The computer device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the sample acquisition method of the driving behavior analysis model in any of the aforementioned embodiments is implemented.

[0157] Instructions are stored in the computer-readable storage medium. When the instructions are executed on the terminal device, the terminal device executes the sample acquisition method of the driving behavior analysis model in any of the aforementioned embodiments.

[0158] It should be noted that each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The device embodiment described above is merely schematic, in which the unit described as a separate component may or may not be physically separated, and the component prompted as a unit may or may not be a physical unit, that is, it may be located in one place, or it may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative work.

[0159] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A sample acquisition method for a driving behavior analysis model, characterized in that: The method comprises: Determine a detection subject and a target conversion vehicle in a first sample, wherein the target conversion vehicle is a first environment vehicle adjacent to the detection subject; Acquire the first posture information of the detection subject and the second posture information of the target conversion vehicle; Acquiring driving behavior data of the detection subject and driving behavior data of the target conversion vehicle; The detection subject and the target conversion vehicle are converted in position according to the first position information and the second position information; the driving behavior data of the detection subject and the target conversion vehicle are converted to obtain a second sample with the target conversion vehicle as a new detection subject.

2. The method according to claim 1, characterized in that The step of performing posture conversion on the detection subject and the target conversion vehicle according to the first posture information and the second posture information specifically includes: The first coordinate system of the detection subject is converted into the coordinate system where the target conversion vehicle is located to obtain a second coordinate system; In the second coordinate system, the fifth posture information of the target conversion vehicle as a new detection subject is obtained according to the coordinate conversion formula and the second posture information; the fourth posture information of the detection subject as an environment vehicle is obtained according to the coordinate conversion formula and the first posture information.

3. The method according to claim 1, characterized in that The acquiring of the driving behavior data of the detection subject and the driving behavior data of the target conversion vehicle specifically includes: Acquire a vehicle driving state of the detection subject, where the vehicle driving state is one of the driving behavior data; Determine a lane line between the detection subject and the target conversion vehicle that is closest to the target conversion vehicle according to the second posture information, and determine a lane line equation corresponding to the lane line; Obtaining a distance between the target conversion vehicle and the lane line according to the second posture information and the lane line equation; The vehicle driving state of the target conversion vehicle is obtained according to the distance.

4. The method according to claim 2, characterized in that: The first sample also includes a second environment vehicle adjacent to the detection subject, and the posture information of the second environment vehicle is third posture information. The method further includes: In the second coordinate system, sixth posture information of the second environment vehicle is obtained according to the coordinate conversion formula and the third posture information.

5. The method according to any one of claims 1 to 4, characterized in that: After obtaining the sample with the target conversion vehicle as the new detection subject, the method further includes: Acquire a first duration that the target conversion vehicle is adjacent to the detection subject; If the first duration is less than a preset threshold, the target conversion vehicle is deleted as a second sample of a new detection subject.

6. The method according to claim 4, characterized in that The method further comprises: Acquire a second duration that the second environment vehicle is adjacent to the detection subject; If the second duration is less than a preset threshold, the second environment vehicle is deleted from the second sample.

7. The method according to claim 1, characterized in that The driving behavior data includes speed and acceleration.

8. A sample acquisition device for a driving behavior analysis model, characterized in that: The device comprises: a determination module, a first acquisition module, a second acquisition module and a conversion module; The determination module is used to determine a detection subject and a target conversion vehicle in a first sample, wherein the target conversion vehicle is a first environment vehicle adjacent to the detection subject; The first acquisition module is used to acquire the first posture information of the detection subject and the second posture information of the target conversion vehicle, The second acquisition module is used to acquire the driving behavior data of the detection subject and the driving behavior data of the target conversion vehicle; The conversion module is used to perform posture conversion on the detection subject and the target conversion vehicle according to the first posture information and the second posture information; and convert the driving behavior data of the detection subject and the target conversion vehicle to obtain a second sample with the target conversion vehicle as the new detection subject.

9. A computer device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the sample acquisition method for the driving behavior analysis model according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed on a terminal device, the terminal device executes the sample acquisition method for the driving behavior analysis model as described in any one of claims 1 to 7.

Images