Method and device for acquiring front vehicle cutting-in scene data

Abstract

The present application provides a method and device for acquiring scene data of a preceding vehicle cutting in, wherein the method includes: acquiring a video file and a corresponding sensor file recorded simultaneously during a historical measurement of a vehicle; establishing a change situation of the target vehicle in the lane line based on the detection and recognition results of the preset features of the target vehicle and the lane line in the video file; calculating a first index when any of the target vehicles is detected to cut in front of the vehicle based on the change situation; calculating a second index when it is detected that the sensor file meets a preset condition; and intercepting data containing the time point of the preceding vehicle cutting in from the video file and the sensor file respectively when the weighted sum of the first index and the second index is greater than a preset threshold. Since the intercepted data containing the time point of the preceding vehicle cutting in can be used for the control logic development of the ADAS controller, the problem of a long development cycle of the control algorithm of the ADAS controller for the preceding vehicle cutting in condition can be solved.

Description

Technical Field

[0001] The present application relates to the field of development and testing of intelligent driving vehicles, and in particular to a method and device for acquiring scene data of a preceding vehicle cutting in. Background Art

[0002] While driving, it's common for a vehicle ahead to cut into a driver's lane from a side lane (referred to as "cut-in" for convenience) and can be quite dangerous in some situations. The Advanced Driving Assistant System (ADAS) controller must handle these situations, requiring the development of control algorithms tailored to these situations.

[0003] The development of ADAS controllers primarily follows a V-shaped model development process. The V-model clearly identifies the various stages of development and testing, with each stage having a clear division of labor, facilitating overall project control. Both software testing and real-vehicle testing are conducted after functional development is complete. Therefore, the parameters required to calculate the ADAS controller's control algorithm for the preceding vehicle's approach must undergo preliminary calibration through software-in-the-loop / hardware-in-the-loop testing. This is followed by a large amount of real-vehicle data, and the final, optimal overall parameters are calibrated based on the test driver's driving habits. This results in a lack of feedback on test results during the early stages of development, and the V-model development process often needs to be repeated to achieve better control over the algorithm and parameters.

[0004] Among them, the later actual vehicle testing and calibration of the front vehicle cutting-in scenario require a larger site and at least two vehicles (the vehicle itself and the vehicle cutting in front of the vehicle). The entire testing process takes a lot of time and money, which in turn leads to a longer development cycle of the ADAS controller's control algorithm for the front vehicle cutting-in condition. Summary of the Invention

[0005] The present application provides a method and device for acquiring preceding vehicle cut-in scenario data, aiming to solve the problem of a long development cycle of the control algorithm of the ADAS controller for the preceding vehicle cut-in condition.

[0006] In order to achieve the above objectives, this application provides the following technical solutions:

[0007] This application provides a method for acquiring preceding vehicle cut-in scene data, comprising:

[0008] Obtain the video files and corresponding sensor files recorded during the vehicle's historical measurement process;

[0009] Establishing a change situation of the target vehicle within the lane line based on the detection and recognition results of the preset features of the target vehicle and the lane line in the video file; the target vehicles include at least: a preset number of vehicles in a lane adjacent to the vehicle and located in front of the vehicle;

[0010] When any of the target vehicles is detected to cut in front of the vehicle according to the change, calculating a first index for representing a probability of a front vehicle cutting in in the video file;

[0011] When it is detected that the sensor file satisfies preset conditions, a second index representing a probability of a preceding vehicle cutting in occurs in the sensor file is calculated; the preset conditions include: a change in the lateral distance between the first vehicle and the vehicle is greater than a preset lateral threshold, a change in the longitudinal distance between the first vehicle and the vehicle is greater than a preset longitudinal threshold, and an identification of the first vehicle changes; the first vehicle being the vehicle that is located in front of the vehicle and closest to the vehicle in its lane;

[0012] When the weighted sum of the first index and the second index is greater than a preset threshold, data containing the time point of the preceding vehicle cutting in is respectively intercepted from the video file and the sensor file.

[0013] Optionally, the preset feature includes a contour;

[0014] The calculating of a first index for representing a probability of a front vehicle cutting in occurs in the video file includes:

[0015] Obtaining an actual value of a first parameter; the first parameter includes at least: clarity and completeness of the contour, and a state of a lane line of a lane in which the vehicle is located;

[0016] Determining the weights corresponding to the actual values ​​of the first parameter according to the preset correspondence between the preset values ​​of the first parameter and the weights;

[0017] The first index is calculated according to the weights corresponding to the actual values ​​of the first parameters.

[0018] Optionally, calculating the first index according to weights corresponding to actual values ​​of the first parameter includes:

[0019] The product of the weights corresponding to the actual values ​​of the first parameter is used as the first index.

[0020] Optionally, the preset features also include: wheel contour; the first parameter also includes: weather conditions, lighting conditions and clarity and completeness of the wheel contour of the target vehicle.

[0021] Optionally, the calculating of the second index for characterizing the probability of a preceding vehicle cutting in on the sensor file includes:

[0022] Obtaining actual values ​​of second parameters from the sensor file; the second parameters including at least: the longitudinal distance change, the longitudinal speed of the vehicle, the steering wheel angle of the vehicle, and the lane status of the vehicle;

[0023] Determining the weights corresponding to the actual values ​​of the second parameter based on a preset correspondence between the preset values ​​of the second parameter and the weights, thereby obtaining the weights corresponding to the actual values ​​of the second parameter;

[0024] The second index is calculated according to the weights corresponding to the actual values ​​of the second parameter.

[0025] Optionally, calculating the second index according to weights corresponding to actual values ​​of the second parameter includes:

[0026] The sum of the weights corresponding to the actual values ​​of the second parameter is used as the second index.

[0027] Optionally, the second parameter also includes: the lateral distance change, the confidence of the lane line, the device that detected the first vehicle, and the type of the first vehicle.

[0028] Optionally, the preset condition also includes:

[0029] The degree of change in the state of the lane line of the lane where the vehicle is located is less than a preset degree threshold, and the lateral distance between the first vehicle and the vehicle changes from large to small before the first vehicle after the logo changes and cuts into the vehicle.

[0030] Optionally, after obtaining the video file and the corresponding sensor file recorded simultaneously with the vehicle's historical time, and before establishing the change status of the target vehicle within the lane line based on the detection and recognition results of the preset features of the target vehicle and the lane line in the video file, the method further includes:

[0031] The video file and the corresponding sensor file are preprocessed; the preprocessing includes: deleting invalid values, processing abnormal values ​​and missing values.

[0032] The present application also provides a device for acquiring a preceding vehicle cut-in scenario database, comprising:

[0033] An acquisition module is used to obtain the video files and corresponding sensor files recorded simultaneously during the vehicle's historical measurement process;

[0034] An establishment module is used to establish a change situation of the target vehicle in the lane line based on the detection and recognition results of the preset features of the target vehicle and the lane line in the video file; the target vehicle includes at least: a preset number of vehicles in the lane beside the vehicle and located in front of the vehicle;

[0035] a first calculation module, configured to calculate a first index representing a probability of a preceding vehicle cutting in in the video file when any of the target vehicles is detected cutting in front of the vehicle according to the change;

[0036] a second calculation module, configured to calculate a second index representing a probability of a preceding vehicle cutting in according to the sensor file when it is detected that the sensor file satisfies preset conditions; the preset conditions comprising: a change in the lateral distance between the first vehicle and the vehicle being greater than a preset lateral threshold, a change in the longitudinal distance between the first vehicle and the vehicle being greater than a preset longitudinal threshold, and a change in the identification of the first vehicle; the first vehicle being the vehicle that is located in front of and closest to the vehicle in the lane of the vehicle;

[0037] The interception module is used to intercept data containing the time point of the preceding vehicle cutting in from the video file and the sensor file respectively when the weighted sum of the first index and the second index is greater than a preset threshold.

[0038] The method and apparatus for acquiring preceding vehicle cut-in scenario data disclosed herein acquire video files and corresponding sensor files recorded simultaneously during historical vehicle measurements; establish a change profile of the target vehicle within the lane lines based on the detection and recognition results of preset features of the target vehicle and lane lines in the video files; and calculate a first index representing the probability of a preceding vehicle cut-in occurring in the video file if any of the target vehicles is detected to have cut in front of the vehicle based on the change profile. A second index representing the probability of a preceding vehicle cut-in occurring in the sensor file is calculated if a change in the lateral distance between the first vehicle and the vehicle exceeds a preset lateral threshold, a change in the longitudinal distance between the first vehicle and the vehicle exceeds a preset longitudinal threshold, and the identification of the first vehicle changes. A second index representing the probability of a preceding vehicle cut-in occurring in the sensor file is calculated. If the weighted sum of the first and second indices exceeds a preset threshold, a preceding vehicle cut-in has occurred. Data containing the time point of the preceding vehicle cut-in is captured from both the video file and the sensor file.

[0039] Since the video file and the corresponding sensor file are recorded during the historical measurement process of the vehicle, the data intercepted by this application containing the time point of the front vehicle's cut-in is the front vehicle's cut-in scene data during the historical measurement process. Therefore, based on the front vehicle's cut-in scene data, the development of the control logic of the ADAS controller for the front vehicle's cut-in condition provides measured data, which can reduce the number of times the V model development process is executed. Furthermore, this application can solve the problem of the long development cycle of the control algorithm of the ADAS controller for the front vehicle's cut-in condition. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. 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 any creative work.

[0041] Figure 1 A schematic diagram of a frame of image captured by a camera disclosed in an embodiment of the present application;

[0042] Figure 2 This is a flow chart of a method for acquiring preceding vehicle cut-in scene data disclosed in an embodiment of the present application;

[0043] Figure 3 A schematic diagram of a calculation process of a first index disclosed in an embodiment of the present application;

[0044] Figure 4 This is a schematic diagram of a typical cutting-in condition of a preceding vehicle disclosed in an embodiment of the present application;

[0045] Figure 5 This is a schematic diagram of a scenario in which a misjudgment of a preceding vehicle cutting in may occur, as disclosed in an embodiment of the present application;

[0046] Figure 6 This is a schematic diagram of a calculation process of a second index disclosed in an embodiment of the present application;

[0047] Figure 7 This is a schematic structural diagram of a device for acquiring scene data of a preceding vehicle cutting in, as disclosed in an embodiment of the present application. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0049] In the embodiments of this application, a camera and radar were installed on the vehicle during historical measurements. The camera must be mounted in a clear, unobstructed position in front of the vehicle, aligned with the driver's line of sight, and capable of clearly capturing information such as lane markings and target vehicles in front of the vehicle. The camera must be able to output stable video and write files to the hard drive according to naming conventions.

[0050] Among them, Figure 1 The image shown in FIG. 1 is used as an example to introduce the meaning of the target vehicle in the embodiment of the present application. In the embodiment of the present application, the target vehicle may include: the vehicle closest to the vehicle in the lane (the lane where the vehicle is located), such as Figure 1 The second closest vehicle in this lane is FT1. Figure 1 FT2 in the left lane of the vehicle. The vehicle closest to the vehicle in longitudinal distance is Figure 1 FT3 in the right lane of the vehicle. The vehicle with the closest longitudinal distance to the vehicle is Figure 1 FT4 in the left lane of the vehicle. The vehicle closest to the vehicle in the lateral direction is as follows: Figure 1 The vehicle in the lane to the right of the vehicle is the closest to the vehicle in the lateral direction, such as Figure 1 FT6 in.

[0051] It should be noted that, in the embodiment of the present application, in the video captured by the camera, each frame of the image not only contains the target vehicle, but may also include other vehicles besides the target vehicle.

[0052] In the embodiment of the present application, the camera outputs the captured video file, and the radar outputs the collected data file (for the convenience of description, referred to as the sensor file). The sensor file includes: information about the target vehicle (including the target vehicle's ID, longitudinal distance (longitudinal distance from the vehicle) / speed / acceleration, lateral distance (lateral distance from the vehicle) / speed / acceleration, attributes (which may include vehicle type) and status, etc.), lane line information of the vehicle lane and the two adjacent lanes (lateral distance (lateral distance from the vehicle), inclination, curvature and rate of change of curvature, lane line type and credibility, etc.), and other information (fences, curbs, etc.).

[0053] In the embodiment of the present application, the vehicle may have conducted multiple field measurements over a historical period. Each field measurement may generate a video file, a sensor file, or a video file and a corresponding sensor file. The generated video file and the corresponding sensor file may be stored in the same storage location with the same name, distinguished only by the file extension.

[0054] Figure 2A method for acquiring preceding vehicle cut-in scene data provided in an embodiment of the present application may include the following steps:

[0055] S201: Obtain video files and corresponding sensor files recorded simultaneously during the historical vehicle measurement process.

[0056] In this step, video files and sensor files are obtained from multiple vehicle measurements over a historical period. Specifically, the video files and corresponding sensor files are extracted, i.e., video files and sensor files are generated during the measurement process in pairs. From these extracted pairs, video file and sensor file pairs are extracted whose difference between the start and end time points falls within a preset difference range.

[0057] In this embodiment, the processing process of any pair of video files and sensor files obtained in this step is the same. For the convenience of description, the following steps are introduced using any pair of video files and sensor files as an example.

[0058] S202: Preprocess the video file and the corresponding sensor file.

[0059] In this step, preprocessing can include: deleting invalid values, processing outliers and missing values.

[0060] It should be noted that, in practice, this step is optional.

[0061] S203: Establishing a change situation of the target vehicle in the lane line based on the detection and recognition results of the preset features of the target vehicle and the lane line in the video file.

[0062] In this step, the target vehicles include at least a preset number of vehicles in the lane beside the vehicle and in front of the vehicle. Figure 1 Of course, in practice, the target vehicle of this step may also include FT1 and FT2. This embodiment does not limit the specific content of the target vehicle.

[0063] In this embodiment, the preset features and lane lines of the target vehicle in each frame of the video file are detected and identified to obtain detection and identification results. The detection and identification results can reflect the changes in lane lines for each target vehicle across multiple consecutive frames of video. Therefore, in this step, the changes in lane lines for the target vehicle can be established. Detection and identification can be implemented using the Faster R-CNN recognition algorithm, where the convolutional neural network (CNN) in the Faster R-CNN recognition algorithm can identify the target vehicle.

[0064] Optionally, in this embodiment, the preset feature may include a contour. Of course, in practice, the preset feature may also include a wheel. This embodiment does not limit the specific content of the preset feature.

[0065] Taking the preset features including contours as an example, in this step, the contours of each target vehicle in the continuous multi-frame images can be associated to obtain the changes in the contours of each target vehicle on the lane line.

[0066] In this step, lane lines, target vehicles in the lane and adjacent left and right lanes, and wheels of the target vehicles can be identified.

[0067] S204 : When any target vehicle is detected cutting in front of the vehicle according to the change situation, a first index representing a probability of the front vehicle cutting in occurs in the video file is calculated.

[0068] In this embodiment, when any target vehicle is detected to have cut in front of the vehicle according to the change situation, it indicates that a front vehicle cut-in condition may have occurred. Therefore, in this step, a first index is calculated to characterize the probability of the front vehicle cutting in in the video file. Figure 3 The corresponding embodiments are introduced and will not be described again here.

[0069] S205 : When it is detected that the sensor file meets a preset condition, a second index representing a probability of a preceding vehicle cutting in occurs in the sensor file is calculated.

[0070] In this embodiment, the vehicle among the pre-defined target vehicles that is located in front of the vehicle and closest to the vehicle in the lane of the vehicle is referred to as the first vehicle.

[0071] In practice, a typical cutting-in condition is as follows: Figure 4 As shown, it can be seen that the lateral distance, longitudinal distance, and identification of the first vehicle to the vehicle (the vehicle itself) change significantly before and after the preceding vehicle cuts in. Therefore, in this embodiment, the preset conditions include at least the first condition, wherein the first condition refers to: the change in the lateral distance between the first vehicle and the vehicle is greater than a preset lateral threshold, the change in the longitudinal distance is greater than a preset longitudinal threshold, and the identification of the first vehicle changes. The preset longitudinal threshold is not less than the length of one vehicle body.

[0072] In this embodiment, the sensor file includes information about the first vehicle, including information about the first vehicle, such as its identification, longitudinal distance (longitudinal distance from the vehicle in question) / speed / acceleration, lateral distance (lateral distance from the vehicle in question) / speed / acceleration, attributes (which may include vehicle type, etc.), and status. Therefore, in this step, based on the information about the first vehicle output by the sensor file, it is possible to determine whether the sensor file satisfies the first condition. If the preset condition is met, it indicates that a preceding vehicle may have cut in. Therefore, in this embodiment, upon detecting that the first condition is met, the second index is calculated.

[0073] In practice, if the first condition is met and it is determined that the front vehicle cuts in, a misjudgment may occur. Figure 5 The misjudgment scenario shown, Figure 5 Among them, (a) shows the scene of the vehicle changing lanes, (b) shows the scene of the vehicle being stationary, (c) shows the scene of the lane line being lost, and (d) shows that the first vehicle has not been stable for a period of time, so these working conditions need to be excluded. In order to improve the accuracy of indicating that the front vehicle cuts in, in practice, when the front vehicle cuts in, the vehicle (the vehicle) is stable in the lane, and in the process of the front vehicle cutting in, the lateral distance between the cutting-in vehicle and the vehicle changes from large to small. Therefore, in this embodiment, the preset conditions may also include a second condition and a third condition, wherein the second condition refers to: the degree of change in the state of the lane line of the lane where the vehicle is located is less than the preset degree threshold. The third condition refers to: before the first vehicle after the change in identification cuts in to the vehicle, the lateral distance between the first vehicle and the vehicle changes from large to small.

[0074] Among them, the value of the state of the lane line of the lane where the vehicle is located in the second condition may include: double-sided lane lines, single-sided lane lines or no lane lines. The degree of change in the state of the lane line of the lane where the vehicle is located refers to: the degree of change in the state of the lane line of the lane where the vehicle is located before and after the moment when the first condition is met in the sensor file. For example, if the state of the lane line changes from double-sided lane lines to no lane lines, the value of the degree of change may be 2; if the state of the lane line changes from double-sided lane lines to single-sided lane lines, the value of the degree of change may be 1. Of course, in practice, how to define the value of the degree of change of the state of the lane line can be determined according to actual conditions, and this embodiment does not limit the specific definition method.

[0075] In this step, the process of calculating the second index is Figure 6 The above is introduced in the embodiment shown and will not be described again here.

[0076] S206 : When the weighted sum of the first index and the second index is greater than a preset threshold, data containing the time point of the preceding vehicle cutting in is respectively intercepted from the video file and the sensor file.

[0077] In this embodiment, if only the first index or the second index is used to determine whether the preceding vehicle has cut in, misjudgment is likely to occur. Therefore, this embodiment combines the first index and the second index to calculate a comprehensive index that represents the probability of the preceding vehicle cutting in. The larger the calculated comprehensive index, the greater the probability of the preceding vehicle cutting in.

[0078] Since the first index Y1 obtained by motion target detection and tracking in the video and the second index Y2 obtained by sensor fusion data processing have been obtained, both indices are dimensionless non-negative values ​​and range from 0 to 1. In this embodiment, the comprehensive index is recorded as Y, where Y = kY1 + (1-k)Y2, and the value of k is between 0 and 1, indicating the weight of Y1.

[0079] To determine effective weight settings, a calibration test can be performed on a small portion of data. For example, a random portion of the preceding vehicle cut-in scenario data can be manually extracted from the dataset. The video and sensor data for this portion of the scenario can be re-injected using different k values. The accuracy rate for each k value can be recorded to determine the correct k value. Once the k value is determined, a comprehensive index for the preceding vehicle cut-in scenario can be obtained for each segment of the re-injected data. When the comprehensive index reaches a preset threshold, it is considered that a preceding vehicle cut-in has occurred.

[0080] In this step, if the comprehensive index exceeds a preset threshold, it indicates that a preceding vehicle has cut in. Therefore, data containing the time points of the preceding vehicle's cut-in is captured from both the video file and the sensor file. The captured data segments can be categorized and saved based on factors such as the road conditions where the preceding vehicle cut in, the type of vehicle that cut in, and the urgency of the cut-in, ultimately creating a library of classified scenarios for preceding vehicle cut-ins.

[0081] In this embodiment, existing real-vehicle test data is leveraged to cost-effectively acquire a database of real-vehicle cut-in scenarios. This scenario database can be used to guide the development of control logic for the ADAS controller's cut-in conditions. Based on the other vehicle's trajectory, it can predict in advance whether the other vehicle will cut into the current lane. This allows the controller to enter the cut-in control mode, activate the control parameters in this mode, and improve control effectiveness, resulting in greater comfort and safety. Furthermore, by comparing the calculation results fed back into the current model with the actual data in the scenario database, subsequent testing and calibration can be reduced, improving efficiency and reducing costs.

[0082] Figure 3 A calculation process of a first index provided in an embodiment of the present application may include the following steps:

[0083] S301: Obtain the actual value of the first parameter.

[0084] In this embodiment, this step is performed when any target vehicle is detected cutting in front of the vehicle. In this step, the first parameter includes at least the clarity and completeness of the target vehicle's outline and the status of the lane markings in the vehicle's lane. In practice, since video signals are also affected by weather, lighting, road conditions, noise, and other factors, the rapid movement of the target vehicle may also cause deformation of the target vehicle's appearance and motion blur. Therefore, in this embodiment, the first parameter may also include weather conditions, lighting conditions, and the clarity and completeness of the target vehicle's wheel outlines.

[0085] S302: Determine the weights corresponding to the actual values ​​of the first parameter according to the preset correspondence between the preset values ​​of the first parameter and the weights.

[0086] In this embodiment, the corresponding relationship between the preset value of the first parameter and the weight is set in advance. Specifically, the corresponding relationship is shown in Table 1.

[0087] Table 1

[0088]

[0089]

[0090] In this embodiment, the weights corresponding to the actual values ​​of the first parameter can be determined based on the correspondence between the preset values ​​of the first parameter and the corresponding weights given in Table 1.

[0091] S303: Calculate a first index according to weights corresponding to actual values ​​of the first parameter.

[0092] Optionally, in this step, calculating the first index based on the weights corresponding to the values ​​of the first parameter may include multiplying the product of the weights corresponding to the values ​​of the first parameter as the first index. Of course, in practice, in addition to multiplying the product of the weights corresponding to the values ​​of the first parameter as the first index, the sum of the weights corresponding to the values ​​of the first parameter may also be used as the first index.

[0093] Figure 6 A calculation process of a second index provided in an embodiment of the present application may include the following steps:

[0094] S601: Obtain an actual value of a second parameter from a sensor file.

[0095] In this embodiment, the second parameter includes at least: a change in the longitudinal distance between the first vehicle and the vehicle, a longitudinal speed of the vehicle, a steering wheel angle of the vehicle, and a lane state of the vehicle.

[0096] In order to improve the accuracy of the calculation result of the second index, in this embodiment, the second parameter may also include: the change in the lateral distance between the first vehicle and the vehicle, the confidence of the lane line, the device that detects the first vehicle and the type of the first vehicle.

[0097] S602: Determine the weights corresponding to the actual values ​​of the second parameter based on the preset correspondence between the preset values ​​of the second parameter and the weights, and obtain the weights corresponding to the actual values ​​of the second parameter.

[0098] In this embodiment, the corresponding relationship between the preset value of the second parameter and the weight is set in advance, as shown in Table 2.

[0099] Table 2

[0100]

[0101] In this step, the weights corresponding to the actual values ​​of the second parameter can be determined based on the correspondence between the preset values ​​of the second parameter and the weights given in Table 2.

[0102] S603: Calculate a second index according to the weights corresponding to the actual values ​​of the second parameter.

[0103] In this step, the sum of the weights corresponding to the actual values ​​of the second parameter can be used as the second index. Of course, in practice, the product of the weights corresponding to the actual values ​​of the second parameter can also be used as the second index. This embodiment does not limit the specific method of calculating the second index based on the weights corresponding to the actual values ​​of the second parameter.

[0104] Figure 7 An apparatus for acquiring scene data of a preceding vehicle cutting in provided in an embodiment of the present application may include: an acquisition module 701, a creation module 702, a first calculation module 703, a second calculation module 704, and an interception module 705, wherein:

[0105] An acquisition module 701 is used to acquire video files and corresponding sensor files recorded simultaneously during the vehicle's historical measurement process;

[0106] Establishing module 702, configured to establish a change situation of the target vehicle within the lane line based on the detection and recognition results of the preset features of the target vehicle and the lane line in the video file; the target vehicles include at least a preset number of vehicles in a lane adjacent to the vehicle and located in front of the vehicle;

[0107] A first calculation module 703 is configured to calculate a first index representing a probability of a preceding vehicle cutting in in the video file when any of the target vehicles is detected cutting in front of the vehicle according to the change condition;

[0108] The second calculation module 704 is configured to calculate a second index representing a probability of a preceding vehicle cutting in according to the sensor file when it is detected that the sensor file satisfies preset conditions; the preset conditions including: a change in the lateral distance between the first vehicle and the vehicle is greater than a preset lateral threshold, a change in the longitudinal distance between the first vehicle and the vehicle is greater than a preset longitudinal threshold, and an identification of the first vehicle changes; the first vehicle being the vehicle that is located in front of and closest to the vehicle in the lane of the vehicle;

[0109] The interception module 705 is configured to intercept data containing a preceding vehicle cut-in time point from the video file and the sensor file respectively when the weighted sum of the first index and the second index is greater than a preset threshold.

[0110] Optionally, the preset feature includes a contour;

[0111] The first calculation module 703 is configured to calculate a first index representing the probability of a preceding vehicle cutting in in the video file, including:

[0112] The first calculation module 703 is specifically used to obtain the actual value of the first parameter; the first parameter includes at least: the clarity and completeness of the contour, and the status of the lane line of the lane where the vehicle is located; based on the preset correspondence between the preset value of the first parameter and the weight, the weight corresponding to the actual value of the first parameter is determined; based on the weight corresponding to the actual value of the first parameter, the first index is calculated.

[0113] Optionally, the first calculation module 703 is configured to calculate the first index according to weights corresponding to actual values ​​of the first parameters, including:

[0114] The first calculation module 703 is specifically configured to multiply the product of the weights corresponding to the actual values ​​of the first parameter as the first index.

[0115] Optionally, the preset features also include: wheel contour; the first parameter also includes: weather conditions, lighting conditions and clarity and completeness of the wheel contour of the target vehicle.

[0116] Optionally, the second calculation module 704 is configured to calculate a second index representing the probability of a preceding vehicle cutting in from the sensor file, including:

[0117] The second calculation module 704 is specifically used to obtain the actual value of the second parameter from the sensor file; the second parameter includes at least: the longitudinal distance change, the longitudinal speed of the vehicle, the steering wheel angle of the vehicle, and the lane line status of the vehicle; based on the preset correspondence between the preset values ​​of the second parameter and the weights, determine the weights corresponding to the actual values ​​of the second parameter, and obtain the weights corresponding to the actual values ​​of the second parameter; calculate the second index based on the weights corresponding to the actual values ​​of the second parameter.

[0118] Optionally, the second calculation module 704 is configured to calculate the second index according to weights corresponding to actual values ​​of the second parameter, including:

[0119] The second calculation module 704 is specifically configured to take the sum of weights corresponding to actual values ​​of the second parameter as the second index.

[0120] Optionally, the second parameter also includes: the lateral distance change, the confidence of the lane line, the device that detected the first vehicle, and the type of the first vehicle.

[0121] Optionally, the preset condition also includes:

[0122] The degree of change in the state of the lane line of the lane where the vehicle is located is less than a preset degree threshold, and the lateral distance between the first vehicle and the vehicle changes from large to small before the first vehicle after the logo changes and cuts into the vehicle.

[0123] Optionally, the device may further include:

[0124] A preprocessing module is used to preprocess the video file and the corresponding sensor file after the acquisition module 701 acquires the video file and the corresponding sensor file recorded simultaneously at the vehicle historical time, and before the establishment module 702 establishes the change situation of the target vehicle in the lane line based on the preset features of the target vehicle and the detection and recognition results of the lane line in the video file; wherein the preprocessing may include: deleting invalid values, processing abnormal values ​​and missing values.

[0125] If the functions described in the method of the embodiment of the present application are implemented in the form of a software functional unit and sold or used as an independent product, they can be stored in a storage medium readable by a computing device. Based on this understanding, the part of the embodiment of the present application that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes a number of instructions for enabling a computing device (which can be a personal computer, server, mobile computing device or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0126] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0127] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for acquiring scene data of a preceding vehicle cutting in, characterized in that: include: Obtain the video files and corresponding sensor files recorded during the vehicle's historical measurement process; Establishing a change situation of the target vehicle in the lane line based on the detection and recognition results of the preset features of the target vehicle and the lane line in the video file; The target vehicles include at least: a preset number of vehicles in a lane beside the vehicle and in front of the vehicle; When any of the target vehicles is detected to cut in front of the vehicle according to the change, calculating a first index for representing a probability of a front vehicle cutting in in the video file; When it is detected that the sensor file satisfies preset conditions, a second index representing a probability of a preceding vehicle cutting in occurs in the sensor file is calculated; the preset conditions include: a change in the lateral distance between the first vehicle and the vehicle is greater than a preset lateral threshold, a change in the longitudinal distance between the first vehicle and the vehicle is greater than a preset longitudinal threshold, and an identification of the first vehicle changes; the first vehicle being the vehicle that is located in front of the vehicle and closest to the vehicle in its lane; When the weighted sum of the first index and the second index is greater than a preset threshold, data containing the time point of the preceding vehicle cutting in is respectively intercepted from the video file and the sensor file.

2. The method according to claim 1, characterized in that The preset features include contours; The calculating of a first index for representing a probability of a front vehicle cutting in occurs in the video file includes: Obtaining an actual value of a first parameter; the first parameter includes at least: clarity and completeness of the contour, and a state of a lane line of a lane in which the vehicle is located; Determining the weights corresponding to the actual values ​​of the first parameter according to the preset correspondence between the preset values ​​of the first parameter and the weights; The first index is calculated according to the weights corresponding to the actual values ​​of the first parameters.

3. The method according to claim 2, characterized in that Calculating the first index according to the weights corresponding to the actual values ​​of the first parameters includes: The product of the weights corresponding to the actual values ​​of the first parameter is used as the first index.

4. The method according to claim 2, characterized in that The preset features also include: wheel contour; the first parameters also include: weather conditions, lighting conditions and the clarity and completeness of the wheel contour of the target vehicle.

5. The method according to claim 1, wherein The calculating of the second index for representing the probability of the preceding vehicle cutting in occurs in the sensor file includes: Obtaining actual values ​​of second parameters from the sensor file; the second parameters including at least: the longitudinal distance change, the longitudinal speed of the vehicle, the steering wheel angle of the vehicle, and the lane status of the vehicle; Determining the weights corresponding to the actual values ​​of the second parameter based on a preset correspondence between the preset values ​​of the second parameter and the weights, thereby obtaining the weights corresponding to the actual values ​​of the second parameter; The second index is calculated according to the weights corresponding to the actual values ​​of the second parameter.

6. The method according to claim 5, characterized in that Calculating the second index according to the weights corresponding to the actual values ​​of the second parameter includes: The sum of the weights corresponding to the actual values ​​of the second parameter is used as the second index.

7. The method according to claim 5, characterized in that The second parameter also includes: the lateral distance change, the confidence level of the lane line, the device that detected the first vehicle, and the type of the first vehicle.

8. The method according to claim 1, characterized in that The preset conditions also include: The degree of change in the state of the lane line of the lane where the vehicle is located is less than a preset degree threshold, and the lateral distance between the first vehicle and the vehicle changes from large to small before the first vehicle after the logo changes and cuts into the vehicle.

9. The method according to claim 1, characterized in that After obtaining the video files and corresponding sensor files simultaneously recorded during the historical vehicle measurement process, and before establishing the change status of the target vehicle in the lane line based on the detection and recognition results of the preset features of the target vehicle and the lane line in the video file, the method further includes: The video file and the corresponding sensor file are preprocessed; the preprocessing includes: deleting invalid values, processing abnormal values ​​and missing values.

10. A device for acquiring a preceding vehicle cutting-in scenario database, characterized in that: include: An acquisition module is used to obtain the video files and corresponding sensor files recorded simultaneously during the vehicle's historical measurement process; An establishment module, configured to establish a change situation of the target vehicle in the lane line based on the detection and recognition results of the preset features of the target vehicle and the lane line in the video file; The target vehicles include at least: a preset number of vehicles in a lane beside the vehicle and in front of the vehicle; a first calculation module, configured to calculate a first index representing a probability of a preceding vehicle cutting in in the video file when any of the target vehicles is detected cutting in front of the vehicle according to the change; a second calculation module, configured to calculate a second index representing a probability of a preceding vehicle cutting in according to the sensor file when it is detected that the sensor file satisfies preset conditions; the preset conditions comprising: a change in the lateral distance between the first vehicle and the vehicle being greater than a preset lateral threshold, a change in the longitudinal distance between the first vehicle and the vehicle being greater than a preset longitudinal threshold, and a change in the identification of the first vehicle; the first vehicle being the vehicle that is located in front of and closest to the vehicle in the lane of the vehicle; The interception module is used to intercept data containing the time point of the preceding vehicle cutting in from the video file and the sensor file respectively when the weighted sum of the first index and the second index is greater than a preset threshold.

Images