Vehicle avoidance scene recognition method and related equipment

Abstract

The invention discloses a vehicle avoidance scene recognition method and device, equipment, a storage medium and a product, and relates to the technical field of intelligent driving, and the vehicle avoidance scene recognition method comprises the steps: determining a driving lane of a vehicle in a historical scene segment under the condition that a vulnerable traffic participant VRU is determined to be a target approaching the vehicle; calculating lane deviation data of the vehicle on the driving lane within a preset time period; based on the lane departure data, vehicle avoidance results are determined, and the vehicle avoidance results include that the vehicle has an avoidance behavior and the vehicle does not have the avoidance behavior. The method provides an effective data basis for improving the decision-making capability of the automatic driving system.

Description

Technical Field

[0001] The present application relates to the field of intelligent driving technology, and in particular to a vehicle avoidance scene recognition method and related equipment. Background Art

[0002] With the development of intelligent driving technology, the demand for data on real traffic scenarios is increasing. After obtaining data on real traffic scenarios, that is, real historical data, if the avoidance scenarios in the historical data can be identified, it can provide a data basis for subsequent training and verification of autonomous driving algorithms, so as to achieve the purpose of improving vehicle perception, decision-making, and control performance.

[0003] However, there is currently a lack of methods to identify whether there are vehicle avoidance VRU scenarios in historical data, resulting in the inability to provide an effective data basis for improving the decision-making capabilities of autonomous driving systems.

[0004] The above content is only used to assist in understanding the technical solution of this application and does not constitute an admission that the above content is prior art. Summary of the Invention

[0005] The main purpose of this application is to provide a vehicle avoidance scene recognition method and related equipment, aiming to solve the technical problem in related technologies that it is difficult to identify whether there are vehicle avoidance VRU scenes in historical data, resulting in the inability to provide an effective data basis for improving the decision-making ability of the autonomous driving system.

[0006] To achieve the above objectives, the present application proposes a vehicle avoidance scene recognition method, which includes:

[0007] When a vulnerable traffic participant (VRU) is determined as a target close to the ego vehicle, a driving lane of the vehicle in the historical scene segment is determined;

[0008] Calculating lane deviation data of the vehicle in the driving lane within a preset time period;

[0009] A vehicle avoidance result is determined based on the lane deviation data, wherein the vehicle avoidance result includes whether the vehicle has an avoidance behavior and whether the vehicle has no avoidance behavior.

[0010] In one embodiment, the step of determining the driving lane of the vehicle in the historical scene segment includes:

[0011] Traversing multiple lane centerlines under the current data frame in the historical scene segment and determining a preliminary associated lane for the vehicle;

[0012] Calculating the projected distance between the vehicle and the center line of each lane;

[0013] A driving lane of the vehicle in the historical scene segment is determined based on the projection distance and the preliminary associated lane.

[0014] In one embodiment, the step of determining the driving lane of the vehicle in the historical scene segment based on the projection distance and the preliminary associated lane includes:

[0015] If the projected distance between the vehicle and the preliminary associated lane is less than or equal to a first preset distance threshold, determining the preliminary associated lane as the driving lane of the vehicle in the historical scene segment;

[0016] If the projected distance between the vehicle and the initially associated lane is greater than a first preset distance threshold, selecting a minimum projected distance from the projected distances corresponding to the center lines of other lanes;

[0017] The lane to which the lane centerline corresponding to the minimum projection distance belongs is determined as the driving lane.

[0018] In one embodiment, the step of calculating lane deviation data of the vehicle in the driving lane within a preset time period includes:

[0019] Determine the distance between the last trajectory point on the current lane centerline and the initial point on the lane centerline;

[0020] Traversing the data frames between the target frame and the preset frame, calculating the longitudinal displacement and lateral displacement of the current position of the vehicle on the lane centerline in each data frame;

[0021] If the longitudinal displacement is greater than or equal to the distance value, or less than a second preset distance threshold, the vehicle is determined to be in front of the lane or at the starting position, and no lane offset data is generated; otherwise, the lateral displacement is added to a preset offset track list;

[0022] Based on the preset deviation trajectory list, lane deviation data of the vehicle in the driving lane is generated.

[0023] In one embodiment, the step of determining a vehicle avoidance result based on the lane deviation data includes:

[0024] Extracting the first displacement point in the preset offset trajectory list as a reference point, wherein the lateral displacement includes multiple displacement points;

[0025] Calculate the lateral offset of each displacement point in the preset displacement trajectory list relative to the reference point;

[0026] Extracting the maximum value of the lateral offsets to obtain the maximum offset value of the lateral displacements;

[0027] If the maximum offset value is greater than a preset offset threshold, it is determined that the vehicle has an avoidance behavior; otherwise, the vehicle has no avoidance behavior.

[0028] In one embodiment, before the step of determining that a vulnerable traffic participant (VRU) is a target close to the vehicle, the method further includes:

[0029] Obtain historical scene fragments during vehicle driving;

[0030] Determine a first trajectory point of the VRU and a second trajectory point of the vehicle corresponding to each data frame in the historical scene segment;

[0031] Based on the first trajectory point and the second trajectory point, it is determined whether the VRU is an object close to the vehicle.

[0032] In one embodiment, the step of determining the first trajectory point of the VRU and the second trajectory point of the vehicle corresponding to each data frame in the historical scene segment includes:

[0033] Traversing the vehicle indexes and object indexes corresponding to the VRUs in each frame from the target frame to the preset frame, wherein the target frame includes the start frame, and the preset frame is separated from the target frame by a preset frame length;

[0034] Based on the object index, the vehicle index, and each of the data frames, a plurality of first trajectory points of the VRU and a plurality of second trajectory points of the vehicle within a preset time period are determined.

[0035] In one embodiment, the step of determining whether the VRU is a target close to the vehicle based on the first trajectory point and the second trajectory point includes:

[0036] calculating a minimum distance between the first trajectory point and the corresponding second trajectory point, and determining displacement data of the VRU relative to the vehicle;

[0037] When it is determined that the displacement data is within a preset range and the minimum distance is less than a preset distance, the VRU is determined to be a target close to the vehicle.

[0038] In addition, to achieve the above-mentioned purpose, the present application also proposes a vehicle avoidance scene recognition device, which includes:

[0039] The first determination module is configured to determine a driving lane of the vehicle in the historical scene segment when a vulnerable traffic participant VRU is determined to be a target close to the vehicle;

[0040] a calculation module, configured to calculate lane deviation data of the vehicle in the driving lane within a preset time period;

[0041] The second determination module is configured to determine a vehicle avoidance result based on the lane deviation data, wherein the vehicle avoidance result includes whether the vehicle has an avoidance behavior or not.

[0042] In addition, to achieve the above-mentioned purpose, the present application also proposes a vehicle avoidance scene recognition device, which includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the computer program is configured to implement the steps of the vehicle avoidance scene recognition method as described above.

[0043] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the steps of the vehicle avoidance scene recognition method as described above.

[0044] The present application proposes a vehicle avoidance scene recognition method, apparatus, equipment, storage medium and computer program product. The present application determines the vehicle's driving lane in historical scene segments by determining that a vulnerable traffic participant VRU is a target approaching the vehicle, calculates the vehicle's lane deviation data in the driving lane within a preset time period, and determines whether the vehicle has an avoidance behavior based on the lane deviation data. By determining the lane deviation of the vehicle in the driving lane, it can be determined whether an avoidance operation is performed when the vulnerable traffic participant VRU is the target approaching the vehicle. Furthermore, the present application itself can accurately identify vehicle avoidance scenes in historical scene segments, providing an effective data basis for improving the decision-making ability of the autonomous driving system. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0046] 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, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0047] Figure 1 A flow chart illustrating a first embodiment of the vehicle avoidance scene recognition method of the present application;

[0048] Figure 2 A flow chart illustrating a second embodiment of the vehicle avoidance scene recognition method of the present application;

[0049] Figure 3A flowchart illustrating a third embodiment of the vehicle avoidance scene recognition method of the present application is provided;

[0050] Figure 4 This is a schematic diagram of the module structure of the vehicle avoidance scene recognition device according to an embodiment of the present application;

[0051] Figure 5 This is a schematic diagram of the device structure of the hardware operating environment involved in the vehicle avoidance scene recognition method in the embodiment of the present application.

[0052] The purpose, features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0053] It should be understood that the specific embodiments described herein are merely used to explain the technical solutions of the present application and are not intended to limit the present application.

[0054] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0055] The main solutions of the embodiments of this application are:

[0056] When a vulnerable traffic participant (VRU) is determined as a target close to the ego vehicle, a driving lane of the vehicle in the historical scene segment is determined;

[0057] Calculating lane deviation data of the vehicle in the driving lane within a preset time period;

[0058] A vehicle avoidance result is determined based on the lane deviation data, wherein the vehicle avoidance result includes whether the vehicle has an avoidance behavior and whether the vehicle has no avoidance behavior.

[0059] Because related technologies often misjudge the vehicle's avoidance maneuvers without the support of historical avoidance data, there is currently a lack of methods that can identify whether there are vehicle avoidance VRU scenarios in the data, resulting in an inability to provide an effective data basis for improving the decision-making capabilities of autonomous driving systems.

[0060] The present application proposes a vehicle avoidance scene recognition method, apparatus, equipment, storage medium and computer program product. The present application determines the vehicle's driving lane in historical scene segments by determining that a vulnerable traffic participant VRU is a target close to the vehicle, calculates the vehicle's lane deviation data in the driving lane within a preset time period, and determines whether the vehicle has an avoidance behavior based on the lane deviation data. By determining the lane deviation of the vehicle in the driving lane, it can be determined whether the vulnerable traffic participant VRU has performed an avoidance operation when the vulnerable traffic participant VRU is a target close to the vehicle. Furthermore, the present application itself can accurately identify vehicle avoidance scenes in historical scene segments, providing an effective data basis for improving the decision-making ability of the autonomous driving system.

[0061] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, mobile phone, etc., or an electronic device capable of implementing the above functions, a vehicle avoidance scene recognition device, etc. The following uses the vehicle avoidance scene recognition device as an example to illustrate this embodiment and the following embodiments.

[0062] Based on this, the embodiment of the present application provides a vehicle avoidance scene recognition method, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the vehicle avoidance scene recognition method of the present application.

[0063] In this embodiment, the vehicle avoidance scene recognition method includes steps S10 to S30:

[0064] Step S10, when the vulnerable traffic participant VRU is determined to be a target close to the vehicle, determining the driving lane of the vehicle in the historical scene segment;

[0065] It should be noted that vehicle avoidance scenarios generally occur when the vehicle needs to avoid certain obstacles. VRU refers to pedestrians, two-wheeled vehicles, three-wheeled vehicles and other relatively disadvantaged parties in traffic. Therefore, it is necessary to perform avoidance operations on VRU during vehicle driving. When it is determined that the disadvantaged traffic participant VRU is a target close to the vehicle, it is necessary to identify whether the vehicle has performed an avoidance operation in the current scenario so that the vehicle can perform the correct action in the next step.

[0066] It should be noted that the vehicle can be a car, a truck, etc., without any specific limitation.

[0067] It should be noted that the historical scene fragments can be multiple data frames within a historical time period. The driving data of the vehicle driving scene are counted under these data frames. The driving data includes various data such as vehicle position, vehicle speed, and VRU position. In the process of determining the vehicle avoidance scene, these data can be comprehensively processed to obtain accurate avoidance scene recognition results.

[0068] It should be noted that the driving lane can be the middle lane, left-turn lane, etc., and there is no specific limitation. Determining the vehicle's driving lane can quickly determine the vehicle's geographical location and contact with pedestrians, and then determine whether the vehicle has performed an avoidance operation based on the changes in the lane to which the vehicle belongs.

[0069] It should be noted that the process of determining the driving lane of the vehicle may be: determining the driving lane of the vehicle in the historical scene segment by determining the position coordinates of the vehicle in the historical scene segment.

[0070] In a feasible implementation manner, the step S10 of determining the driving lane of the vehicle in the historical scene segment includes:

[0071] Traversing multiple lane centerlines under the current data frame in the historical scene segment and determining a preliminary associated lane for the vehicle;

[0072] It should be noted that the preliminary associated lane of the vehicle needs to be determined first. The preliminary associated lane is the lane where the vehicle is located in the starting frame. The preliminary associated lane can be determined based on the position coordinates of the vehicle.

[0073] It should be noted that there is a lane centerline in the middle of each lane. Since there is not only one lane on the road, the number of lanes must be greater than or equal to 2. Therefore, it is necessary to traverse the lane centerlines that appear in the current data frame to determine the actual number of lanes in the data frame.

[0074] It should be noted that after traversing each lane centerline, the vehicle's preliminary associated lane, as well as the associated lane projection type and minimum projection distance can be obtained.

[0075] Calculating the projected distance between the vehicle and the center line of each lane;

[0076] It should be noted that the projection distance is the distance value between the vehicle and the center line connecting each lane. The projection distance can be 0, 2, 3, etc., and is not specifically limited.

[0077] For example, when a vehicle is in lane 1, the projected distance of the vehicle relative to lane 1 is 0. When lane 2 is adjacent to lane 1, the projected distance of the vehicle relative to lane 1 is 1 or 2. Similarly, the farther the lane centerline is from the vehicle, the longer the projected distance will be, which means that the correlation between the lane and the vehicle will be weaker.

[0078] A driving lane of the vehicle in the historical scene segment is determined based on the projection distance and the preliminary associated lane.

[0079] It should be noted that the vehicle's driving lane can be determined based on the projection distance and the preliminary associated lane. This driving lane is for the current data frame. For historical scene fragments, the subsequent driving lane may change. The lane in which the vehicle is traveling can be determined based on the length of the projection distance and the position of the preliminary associated lane.

[0080] In a feasible implementation manner, the step of determining the driving lane of the vehicle in the historical scene segment based on the projection distance and the preliminary associated lane includes:

[0081] If the projected distance between the vehicle and the preliminary associated lane is less than or equal to a first preset distance threshold, determining the preliminary associated lane as the driving lane of the vehicle in the historical scene segment;

[0082] It should be noted that the first preset distance threshold may be 2, 3, 4, etc., and may be adjusted according to the width of the lane, without specific limitation.

[0083] It should be noted that, taking the first preset distance threshold equal to 3 as an example, if the projected distance between the vehicle and the preliminary associated lane is less than or equal to 3, it means that the lane in which the vehicle is located is the preliminary associated lane. Therefore, the preliminary associated lane is determined to be the lane in which the vehicle was traveling in the historical scene segment.

[0084] If the projected distance between the vehicle and the initially associated lane is greater than a first preset distance threshold, selecting a minimum projected distance from the projected distances corresponding to the center lines of other lanes;

[0085] It should be noted that when the projection distance between the vehicle and the initially associated lane is greater than the first preset distance threshold, it means that the lane corresponding to the vehicle is another lane. The minimum projection distance can be selected from the projection distances corresponding to the center lines of other lanes, and the lane corresponding to the minimum projection distance is the driving lane.

[0086] The lane to which the lane centerline corresponding to the minimum projection distance belongs is determined as the driving lane.

[0087] It should be noted that the lane to which the lane centerline corresponding to the minimum projection distance belongs is selected as the driving lane. When the projection distance of the vehicle on the centerline of other lanes has a minimum value, the minimum projection distance shall be used as the standard. In this way, the driving lane in which the vehicle is located can be accurately identified.

[0088] It should be noted that if the minimum projection distance of the ego vehicle in the frame is less than or equal to 3, the ego vehicle's lane association is valid, and the lane bound to the ego vehicle in the data frame is the preliminary lane association. If distance_small is greater than 3, the ego vehicle's lane association is invalid, and the ego vehicle's lane bound to the frame is marked as -1, and the other lanes are set as the ego vehicle's driving lane.

[0089] Step S20, calculating lane deviation data of the vehicle in the driving lane within a preset time period;

[0090] It should be noted that the preset time period may be 5 minutes, 10 minutes, etc. The preset time period corresponds to the sum of the time of each data frame and is not specifically limited.

[0091] It should be noted that the vehicle is always in a moving state in different data frames. Therefore, when in different data, the position of the vehicle is also different. It is necessary to calculate the vehicle's offset data to determine whether the vehicle has performed an avoidance operation.

[0092] It should be noted that the lane offset data can be the longitudinal displacement or lateral displacement of the vehicle. The specific offset amount shall be based on the actual calculation result. Whether the vehicle is in an avoidance scenario is determined based on the longitudinal displacement or lateral displacement of the vehicle.

[0093] Step S30: determining a vehicle avoidance result based on the lane deviation data, wherein the vehicle avoidance result includes whether the vehicle has an avoidance behavior or not.

[0094] It should be noted that when determining the longitudinal displacement or lateral displacement of the lane, the lateral displacement is compared with a preset threshold to determine whether the vehicle has an avoidance behavior. This is mainly to determine whether a lateral detour occurs when the VRU is a target close to the vehicle. If a lateral detour occurs, it is determined that the vehicle has an avoidance behavior. Otherwise, the vehicle does not have an avoidance behavior.

[0095] The present application proposes a vehicle avoidance scene recognition method, apparatus, equipment, storage medium and computer program product. The present application determines the vehicle's driving lane in historical scene segments by determining that a vulnerable traffic participant VRU is a target close to the vehicle, calculates the vehicle's lane deviation data in the driving lane within a preset time period, and determines whether the vehicle has an avoidance behavior based on the lane deviation data. By determining the lane deviation of the vehicle in the driving lane, it can be determined whether the vulnerable traffic participant VRU has performed an avoidance operation when the vulnerable traffic participant VRU is a target close to the vehicle. Furthermore, the present application itself can accurately identify vehicle avoidance scenes in historical scene segments, providing an effective data basis for improving the decision-making ability of the autonomous driving system.

[0096] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above embodiment 1 can be referred to the above introduction and will not be described in detail later. Figure 2 The step S20 of calculating the lane deviation data of the vehicle in the driving lane within a preset time period includes:

[0097] Step S21, determining the distance between the last trajectory point on the current lane centerline and the initial point of the lane centerline;

[0098] It should be noted that the lane centerline is composed of multiple trajectory points. The trajectory points are randomly selected from the front to the back of the lane centerline. The initial point of the lane centerline is the first trajectory point, and the distance between the last trajectory point and the initial point represents the length of the lane centerline.

[0099] It should be noted that the distance value can be 10m, 20m, etc., and is not specifically limited.

[0100] It should be noted that, in addition to the distance value representation, the coordinate value of the trajectory point can also be used to determine whether the vehicle's coordinates are at the initial point of the lane or in front of the lane.

[0101] Step S22, traversing the data frames between the target frame and the preset frame, and calculating the longitudinal displacement and lateral displacement of the current position of the vehicle on the lane centerline in each data frame;

[0102] It should be noted that the target frame may be the initial frame, the second frame, the fifth frame, etc., without any specific limitation.

[0103] It should be noted that the preset frame can be a data frame 60 frames after the target frame, or a data frame 50 frames after the target frame. There is no specific limitation. The number of interval frames is determined by the maximum search time and can be changed as needed.

[0104] It should be noted that the longitudinal displacement may be the displacement value of the vehicle along the center line of the lane, and the lateral displacement may be the displacement value of the vehicle in the direction perpendicular to the center line of the lane.

[0105] Specifically, the longitudinal displacement and lateral displacement of the current position of the vehicle on the lane centerline in each data frame may be calculated as follows:

[0106] Loop through the data of 60 time frames starting from the target frame with a step size of 5 frames, that is, read data every 5 frames. For each sampling frame:

[0107] Check whether its frame index exceeds the frame length of the data segment data_vector. If so, exit the loop early.

[0108] Extract the coordinates of the ego vehicle in the sampling frame (ego_x, ego_y);

[0109] Calculate the s and l values ​​of the ego vehicle’s current position (ego_x, ego_y) in the lane centerline, denoted as (ego_curr_s, ego_curr_l) (s represents the distance along the lane, l represents the distance perpendicular to the lane).

[0110] The longitudinal and lateral displacements of the vehicle are determined based on the difference between the vehicle's position coordinates and the coordinates of the lane centerline.

[0111] Step S23: If the longitudinal displacement is greater than or equal to the distance value, or less than a second preset distance threshold, the vehicle is determined to be in the front of the lane or at the starting position, and no lane offset data is generated; otherwise, the lateral displacement is added to a preset offset trajectory list;

[0112] It should be noted that the second preset distance threshold may be 1e-9 or other values, which are not specifically limited.

[0113] It should be noted that the distance value is the distance between the last trajectory point and the initial point. If the longitudinal displacement is greater than or equal to the distance value, it means that the vehicle is already outside the lane line, or is at the starting point or end point of the lane. At this time, the lateral displacement has no reference value and can be skipped without processing.

[0114] It should be noted that when the vehicle's longitudinal displacement does not meet these two conditions, it means that the vehicle is inside the driving lane, and the lateral displacement can be added to the preset offset trajectory list. Subsequently, the vehicle's lateral displacement is mainly used to determine whether the vehicle has performed an avoidance operation. The preset offset trajectory list includes the vehicle's lateral displacement in different data frames.

[0115] Step S24: generating lane deviation data of the vehicle in the driving lane based on the preset deviation trajectory list.

[0116] It should be noted that after the lateral displacement of the vehicle is counted using the preset offset trajectory list, the lane offset data of the vehicle is generated. At this time, the lane offset data mainly includes the lateral displacement of the vehicle.

[0117] In a feasible implementation manner, the step S30 of determining the vehicle avoidance result based on the lane deviation data includes:

[0118] Extracting the first displacement point in the preset offset trajectory list as a reference point, wherein the lateral displacement includes multiple displacement points;

[0119] It should be noted that the displacement point can be the position of the vehicle after performing the displacement operation. The first point is selected as the reference point. When counting the displacement points of the vehicle, the first point is the initial point of the vehicle. By comparing the initial point with other displacement points, the degree of deviation of the vehicle position can be determined.

[0120] Calculate the lateral offset of each displacement point in the preset displacement trajectory list relative to the reference point;

[0121] It should be noted that the lateral offset is the lateral offset value of each displacement point relative to the reference point, wherein the lateral offset value can be 1m, 3m, etc., and is not specifically limited.

[0122] It should be noted that the lateral offset may be calculated by calculating the offset difference between each displacement point and the reference point to obtain the lateral offset.

[0123] Extracting the maximum value of the lateral offsets to obtain the maximum offset value of the lateral displacements;

[0124] It should be noted that among the multiple lateral offsets, there generally exists a maximum value and a minimum value. The maximum value among these lateral offsets is taken as the maximum offset value in the lateral displacement.

[0125] If the maximum offset value is greater than a preset offset threshold, it is determined that the vehicle has an avoidance behavior; otherwise, the vehicle has no avoidance behavior.

[0126] It should be noted that the preset offset threshold may be 0.5m, 1m, etc., and is not specifically limited.

[0127] It should be noted that when the maximum offset value is greater than the preset offset threshold, it indicates that at least one point in the trajectory has a lateral offset exceeding the preset offset threshold, indicating that lateral avoidance behavior is in effect. Otherwise, if the lateral offset of all points does not exceed the preset offset threshold, lateral avoidance behavior is not in effect.

[0128] Specifically, to determine whether there is a lateral detour in the preset offset trajectory list, the main idea is to detect whether the lateral offset (yaw change) of a given point relative to the initial point exceeds the preset threshold. The specific judgment method is as follows:

[0129] If the length of the preset offset trajectory list is less than 2, that is, it contains only one point or is empty, no judgment is performed because at least two points are required to define a change trend.

[0130] Define a variable (preset offset threshold) to represent the lateral offset threshold, with a default value of 0.5. This value can be adjusted according to the specific application scenario.

[0131] Take the first point from the list of preset offset tracks as the reference point.

[0132] For each displacement point in the preset offset trajectory list, calculate its lateral offset relative to the reference point to obtain a new list rel_l. Further calculate the maximum absolute value of all elements in rel_l, which represents the maximum lateral offset in the trajectory.

[0133] If the maximum lateral deviation is greater than the preset deviation threshold, it means that the lateral deviation of at least one point in the trajectory exceeds the preset deviation threshold, indicating that lateral avoidance behavior is performed. Otherwise, if the lateral deviation of all points does not exceed the preset deviation threshold, it indicates that lateral avoidance behavior is not performed.

[0134] Furthermore, the calculated offset data can be used to further determine whether a VRU avoidance scenario occurs. For example, if the vehicle's lateral avoidance behavior occurs in the same frame, and the amount of VRU data surrounding the vehicle is greater than N (depending on the actual setting), the frame can be considered to have experienced VRU avoidance. Furthermore, the proportion of frames with VRU avoidance behavior can be set to indicate whether the data segment represents a VRU avoidance scenario.

[0135] In this embodiment, by determining the vehicle's driving lane and whether the vehicle has lateral displacement, it is possible to accurately calculate whether the vehicle has performed an avoidance maneuver, and further determine whether the vehicle has performed an avoidance behavior.

[0136] Based on the first embodiment and / or the second embodiment of the present application, in the third embodiment of the present application, the same or similar contents as those in the first and second embodiments above can be referred to the above introduction and will not be described in detail later. Figure 3 Before step S10 in which the vulnerable traffic participant VRU is determined to be a target close to the vehicle, the method further includes:

[0137] Step S100, obtaining historical scene segments during vehicle driving;

[0138] It should be noted that the historical scene fragments can be multiple data frames collected within a historical time period during the vehicle's driving process, which include various elements such as vehicles, pedestrians, and road signs; among them, the historical time period can be 10 minutes, 20 minutes, 30 minutes, etc., and there is no specific limitation.

[0139] Step S110, determining a first trajectory point of the VRU and a second trajectory point of the vehicle corresponding to each data frame in the historical scene segment;

[0140] It should be noted that VRU types can include pedestrians, two-wheeled vehicles, three-wheeled vehicles, and so on. Each VRU type is stored in the object attribute category list VRU_type_list. This list also contains a corresponding VRU index list, which stores the object indexes belonging to the VRU in the frame. By traversing all objects in the frame, the indexes of objects whose object category attributes are in the list VRU_type_list are stored in the VRU index list.

[0141] It should be noted that a counter is also predefined, wherein the counter is initialized to 0 and is used to count the specific number of VRUs in the current data frame. When it is determined that the number of VRUs in the current data frame is 5, the count of the counter is also 5.

[0142] It should be noted that for each VRU target, when a vulnerable traffic participant VRU is detected in a historical scene segment, the VRU in each data frame in the scene segment and the driving trajectory of the current vehicle are judged to determine whether the VRU object is in a state close to the vehicle.

[0143] It should be noted that the first trajectory point is a motion trajectory point of the VRU, and the first moving trajectory may be a motion trajectory of the VRU obtained based on the first trajectory point. The first moving trajectory is determined by calculating the first trajectory point of the VRU under different data frames.

[0144] It should be noted that the second trajectory point is the motion trajectory point of the vehicle, and the second moving trajectory can be a motion trajectory obtained based on the second trajectory point of the vehicle. The vehicle is based on the vehicle itself. For each data frame, the vehicle will have a corresponding position coordinate, and then determine a motion trajectory point. Based on these trajectory points, the second moving trajectory is generated.

[0145] Step S120: Determine whether the VRU is a target close to the vehicle based on the first trajectory point and the second trajectory point.

[0146] It should be noted that in the related art, the proximity between the VRU and the vehicle is determined by the relative position of the VRU and the vehicle. The relative position is basically used to determine whether the two are close to each other based on the position in a certain frame. However, this method will cause inaccurate recognition because the amount of data obtained is too small. Even if the amount of data obtained is large enough, there will be errors in the judgment based on the position, and thus, it will be impossible to accurately identify whether the VRU is a target close to the vehicle.

[0147] It should be noted that by determining whether the VRU is a target close to the vehicle based on the first trajectory point and the second trajectory point, more specifically, it can be determined whether the VRU is a target close to the vehicle based on the first travel trajectory and the second travel trajectory. From the perspective of the determination method, it is more intuitive and can directly determine the proximity between the vehicle and the VRU at all times. Therefore, the identification method is more accurate and there will be no false alarms or missed alarms.

[0148] It should be noted that the specific method of determining the proximity between the VRU and the vehicle through the trajectory points of the two can be to determine the distance between the trajectory points on the first moving trajectory and the second moving trajectory based on the distribution of the motion trajectory. When the distance is less than a certain threshold, it can be determined that the two are in a proximity state. Otherwise, they are not in a proximity state, and there is no need to issue an early warning.

[0149] In a feasible implementation manner, the step S110 of determining the first travel trajectory of the VRU and the second travel trajectory of the vehicle corresponding to each data frame in the historical scene segment includes:

[0150] Traversing the vehicle indexes and object indexes corresponding to the VRUs in each frame from the target frame to the preset frame, wherein the target frame includes the start frame, and the preset frame is separated from the target frame by a preset frame length;

[0151] It should be noted that the target frame may be the first frame, the second frame, the fifth frame, etc., and the preset frame may be the 30th frame, the 40th frame, etc., without specific limitation.

[0152] It should be noted that the first preset frame number can be 40 frames, 60 frames, 70 frames, etc., and is not specifically limited.

[0153] It should be noted that the vehicle index is the index value corresponding to the ego vehicle. Similarly, the object index is the index value corresponding to the VRU. In each data frame, there is a frame index, vehicle index, and object index. These index values ​​are used to calculate the motion trajectory of the VRU or the ego vehicle.

[0154] Determining, based on the object index, the vehicle index, and each of the data frames, a plurality of first trajectory points of the VRU and a plurality of second trajectory points of the vehicle within a preset time period;

[0155] It should be noted that the preset time period can be half an hour, one hour, etc., and there is no specific limitation.

[0156] It should be noted that the first trajectory point is a trajectory point calculated by the VRU in each data frame. Similarly, the second trajectory point is a trajectory point calculated by the vehicle in each data frame.

[0157] It should be noted that for each data frame, the object index and vehicle index can be used to determine the trajectory point of the vehicle or VRU in the current data frame. The calculation method of the vehicle and VRU trajectory points is similar. Here, the trajectory point calculation process of the VRU object is used as an example:

[0158] The details are as follows:

[0159] 1. Initialize an empty list to store the object's trajectory points.

[0160] 2. Define the maximum number of steps as 60, which means that the data of up to 60 frames forward can be judged.

[0161] 3. Use a loop to iterate through future frames starting from the current frame, with a step size of 5, meaning data is read every five frames. For each frame, check whether its frame index exceeds the frame length of the data segment. If so, exit the loop early. Also check whether the current frame contains the specified VRU object index. If the VRU object does not exist, exit the loop early. If the VRU object index exists, obtain the VRU object's location information and add it to the preset trajectory list. This provides the location information for both the VRU and the vehicle, allowing the trajectory points of both to be determined.

[0162] A first moving trajectory is generated based on each of the first trajectory points, and a second moving trajectory is generated based on each of the second trajectory points.

[0163] It should be noted that after determining the first trajectory points in each data frame, the set formed by the first trajectory points is connected to obtain the first moving trajectory. Similarly, the second trajectory points are connected to generate the second moving trajectory.

[0164] In a feasible implementation manner, the step S120 of determining whether the VRU is a target close to the vehicle based on the first trajectory point and the second trajectory point includes:

[0165] calculating a minimum distance between the first trajectory point and the corresponding second trajectory point, and determining displacement data of the VRU relative to the vehicle;

[0166] It should be noted that to determine whether the VRU is a target close to the vehicle, it is necessary to calculate the minimum distance between the first trajectory point and the corresponding second trajectory point. Specifically, it is necessary to calculate the minimum distance between the two travel trajectories. When the minimum distance is less than a certain degree / certain threshold, it can be determined whether the VRU is a target close to the vehicle.

[0167] It should be noted that the VRU may have corresponding displacement data relative to the vehicle. The displacement data includes longitudinal displacement and lateral displacement. If the proximity is determined based solely on the movement trajectories of the two, errors may occur. Therefore, it is necessary to determine the displacement data of the VRU relative to the vehicle to determine whether the VRU has the possibility of approaching the vehicle under different data frames.

[0168] In a feasible implementation, the step of calculating the minimum distance between the trajectory points of the first moving trajectory and the trajectory points of the second moving trajectory includes:

[0169] Determining a first trajectory length of the VRU corresponding to a first travel trajectory based on the first trajectory point, and determining a second trajectory length of the vehicle corresponding to a second travel trajectory based on the second trajectory point;

[0170] It should be noted that the first trajectory length is the length of the VRU's trajectory, which is used to represent the length of the VRU's movement. Similarly, the second trajectory length is used to represent the length of the vehicle's movement.

[0171] generating a trajectory point matrix based on the first trajectory length and the second trajectory length, wherein each element of the trajectory point matrix represents a Euclidean distance between any trajectory points on two travel trajectories;

[0172] It should be noted that the process of generating the trajectory point matrix is ​​as follows:

[0173] The length of the first trajectory is calculated as M, the length of the second trajectory is calculated as N, and the distance between any point on the first trajectory and any point on the second trajectory is calculated to obtain an M*N matrix, that is, a trajectory point matrix, where the element (i, j) represents the Euclidean distance between the i-th point on the first trajectory and the j-th point on the second trajectory.

[0174] Based on the trajectory point matrix, a minimum distance between the first trajectory point and the corresponding second trajectory point is calculated.

[0175] It should be noted that finding the minimum value in the M*N matrix is ​​the minimum distance between the first moving track and the second moving track.

[0176] In one possible implementation, the step of determining the displacement data of the VRU relative to the vehicle further includes:

[0177] Obtain VRU coordinate information, vehicle coordinate information, and vehicle heading angle in each data frame;

[0178] It should be noted that the vehicle heading angle can be 30 degrees, 60 degrees, etc., and there is no specific limitation.

[0179] It should be noted that in each data frame, the VRU coordinate information, vehicle coordinate information and vehicle orientation angle can be extracted through the collected scene segments, and then the displacement data of the VRU relative to the vehicle can be calculated based on these data.

[0180] Specifically, according to the specified frame index, the ego vehicle information of the frame is extracted, including the X coordinate (ego_x), Y coordinate (ego_y) and orientation angle (ego_theta), as well as the coordinates of the VRU object index (global_x, global_y), and the longitudinal displacement and lateral displacement of the VRU object index relative to the ego vehicle position in the ego vehicle coordinate system are calculated.

[0181] Calculating a coordinate offset value between the VRU coordinate information and the vehicle coordinate information;

[0182] It should be noted that the coordinate offset value is the offset value of the VRU coordinate information relative to the vehicle coordinate information. The coordinate offset value includes the lateral offset value and the longitudinal offset value. The calculation formula is as follows:

[0183] Longitudinal offset value: x_diff = global_x - ego_x

[0184] Horizontal offset value: y_diff = global_y - ego_y

[0185] Based on the coordinate offset value and the vehicle heading angle, the displacement data of the VRU relative to the vehicle is calculated.

[0186] It should be noted that the calculation formula for the displacement data of the VRU relative to the vehicle based on the coordinate offset value and the vehicle heading angle is as follows:

[0187] angle_cos=math.cos(ego_theta) represents the cosine value of the vehicle's heading angle.

[0188] angle_sin = math.sin(ego_theta), which represents the sin value of the vehicle's heading angle

[0189] local_x=x_diff*angle_cos-y_diff*angle_sin. Multiply the cosine value of the vehicle's heading angle by the longitudinal offset value, and then subtract the product of the lateral offset value and the sin value of the vehicle's heading angle to obtain the longitudinal displacement of the VRU relative to the vehicle.

[0190] local_y=x_diff*angle_sin+y_diff*angle_cos. Multiply the cosine value of the vehicle's heading angle by the lateral offset value, and then add the product of the longitudinal offset value and the sin value of the vehicle's heading angle to obtain the lateral displacement of the VRU relative to the vehicle.

[0191] It should be noted that when determining the displacement of a VRU, it can be determined by longitudinal displacement or by using two types of displacement data. In the actual displacement process, longitudinal displacement is easier to determine whether the two are in a close state, so longitudinal displacement can be used preferentially for determining the close state.

[0192] If it is determined that the displacement data is within a preset range and the minimum distance is less than the preset distance, the VRU is determined to be a target close to the vehicle; otherwise, the VRU is not a target close to the vehicle.

[0193] It should be noted that the preset range may be (0, 30m), (5m, 40m), etc., and is not specifically limited.

[0194] It should be noted that the preset distance can be 1.5m, 3m, etc., and there is no specific limitation.

[0195] It should be noted that the preset range is (0, 30m) and the preset distance is 1.5m. local_x represents the displacement data. If (1) local_x>0, (2) local_x<30, and (3) distance<1.5 are satisfied, the VRU object can be considered as a target in the state of approaching the ego vehicle. Otherwise, the VRU object is not considered as a target in the state of approaching the ego vehicle.

[0196] In this embodiment, whether the VRU is a target that the ego vehicle is approaching is determined by the minimum distance between the two moving trajectories and the displacement data of the VRU relative to the ego vehicle, so as to improve the accuracy of the recognition result and reduce the occurrence of misjudgment and missed judgment.

[0197] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the vehicle avoidance scene recognition method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0198] This application also provides a vehicle avoidance scene recognition device, please refer to Figure 4, the vehicle avoidance scene recognition device includes:

[0199] A first determination module 10 is configured to determine a driving lane of the vehicle in the historical scene segment when a vulnerable traffic participant VRU is determined to be a target close to the vehicle;

[0200] A calculation module 20, configured to calculate lane deviation data of the vehicle in the driving lane within a preset time period;

[0201] The second determining module 30 is configured to determine a vehicle avoidance result based on the lane deviation data, wherein the vehicle avoidance result includes whether the vehicle has an avoidance action or not.

[0202] Optionally, the first determining module includes:

[0203] a first traversal unit, configured to traverse a plurality of lane centerlines under a current data frame in the historical scene segment and determine a preliminary associated lane of the vehicle;

[0204] A first calculation unit is used to calculate the projected distance between the vehicle and the center line of each lane;

[0205] The first determining unit is configured to determine the driving lane of the vehicle in the historical scene segment based on the projection distance and the preliminary associated lane.

[0206] Optionally, the determining unit includes:

[0207] a first determining subunit, configured to determine, if a projected distance between the vehicle and the preliminary associated lane is less than or equal to a first preset distance threshold, that the preliminary associated lane is the driving lane of the vehicle in the historical scene segment;

[0208] a selection subunit, configured to select a minimum projected distance from the projected distances corresponding to the center lines of other lanes if the projected distance between the vehicle and the preliminarily associated lane is greater than a first preset distance threshold;

[0209] The second determining subunit is configured to determine that the lane to which the lane centerline corresponding to the minimum projection distance belongs is the driving lane.

[0210] Optionally, the calculation module includes:

[0211] A second determining unit is used to determine the distance between the last trajectory point on the current lane centerline and the initial point of the lane centerline;

[0212] A second traversal unit is configured to traverse the data frames between the target frame and the preset frame, and calculate the longitudinal displacement and the lateral displacement of the current position of the vehicle on the lane centerline in each of the data frames;

[0213] an adding unit, configured to determine that the vehicle is in front of the lane or at the starting position if the longitudinal displacement is greater than or equal to the distance value, or less than a second preset distance threshold, and not generate lane offset data; otherwise, add the lateral displacement to a preset offset trajectory list;

[0214] The first generating unit is configured to generate lane deviation data of the vehicle in the driving lane based on the preset deviation trajectory list.

[0215] Optionally, the second determining module includes:

[0216] a first extraction unit, configured to extract a first displacement point in a preset offset trajectory list as a reference point, wherein the lateral displacement includes a plurality of displacement points;

[0217] a second calculation unit, configured to calculate a lateral offset of each displacement point in the preset displacement trajectory list relative to the reference point;

[0218] a second extraction unit, configured to extract a maximum value among the lateral offsets to obtain a maximum offset value among the lateral displacements;

[0219] The third determining unit is configured to determine that the vehicle has an avoidance behavior if the maximum offset value is greater than a preset offset threshold, and otherwise, the vehicle has no avoidance behavior.

[0220] Optionally, the device further comprises:

[0221] An acquisition module is used to obtain historical scene fragments during the vehicle's driving process;

[0222] A third determining module is used to determine a first trajectory point of the VRU and a second trajectory point of the vehicle corresponding to each data frame in the historical scene segment;

[0223] A fourth determining module is configured to determine whether the VRU is a target close to the vehicle based on the first trajectory point and the second trajectory point.

[0224] Optionally, the third determining module includes:

[0225] A third traversal unit is configured to traverse the vehicle indexes and object indexes corresponding to the VRUs in each frame from a target frame to a preset frame, wherein the target frame includes a start frame, and the preset frame is separated from the target frame by a preset frame length;

[0226] The fourth determining unit is configured to determine a plurality of first trajectory points of the VRU and a plurality of second trajectory points of the vehicle within a preset time period based on the object index, the vehicle index, and each of the data frames.

[0227] Optionally, the fourth determining module includes:

[0228] a third calculation unit, configured to calculate a minimum distance between the first trajectory point and the corresponding second trajectory point, and determine displacement data of the VRU relative to the vehicle;

[0229] The fifth determining unit is configured to determine that the VRU is a target close to the vehicle when it is determined that the displacement data is within a preset range and the minimum distance is less than a preset distance.

[0230] The vehicle avoidance scene recognition device provided in this application utilizes the vehicle avoidance scene recognition method described in the aforementioned embodiments to address the technical issues surrounding vehicle avoidance scene recognition. Compared to the prior art, the vehicle avoidance scene recognition device provided in this application achieves the same beneficial effects as the vehicle avoidance scene recognition method described in the aforementioned embodiments. Other technical features of the vehicle avoidance scene recognition device are the same as those disclosed in the aforementioned embodiments and are not further elaborated here.

[0231] The present application provides a vehicle avoidance scene recognition device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the vehicle avoidance scene recognition method in the above-mentioned embodiment one.

[0232] Reference below Figure 5 , which shows a schematic structural diagram of a vehicle avoidance scenario recognition device suitable for implementing embodiments of the present application. The vehicle avoidance scenario recognition device in embodiments of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 5 The vehicle avoidance scene recognition device shown is merely an example and should not impose any limitations on the functions and scope of use of the embodiments of the present application.

[0233] like Figure 5As shown, the vehicle avoidance scene recognition device may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. Various programs and data required for the operation of the vehicle avoidance scene recognition device are also stored in RAM 1004. The processing device 1001, ROM 1002, and RAM 1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the vehicle avoidance scene recognition device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows a vehicle avoidance scene recognition device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems can be implemented or have instead.

[0234] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0235] The vehicle avoidance scene recognition device provided in this application utilizes the vehicle avoidance scene recognition method described in the aforementioned embodiment to address the technical issues surrounding vehicle avoidance scene recognition. Compared to the prior art, the vehicle avoidance scene recognition device provided in this application achieves the same beneficial effects as the vehicle avoidance scene recognition method described in the aforementioned embodiment. Other technical features of the vehicle avoidance scene recognition device are the same as those disclosed in the aforementioned embodiment and are not further elaborated here.

[0236] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0237] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0238] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, a computer program) stored thereon, and the computer-readable program instructions are used to execute the vehicle avoidance scene recognition method in the above-mentioned embodiment.

[0239] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0240] The computer-readable storage medium may be included in the vehicle avoidance scene recognition device; or may exist independently without being assembled into the vehicle avoidance scene recognition device.

[0241] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the vehicle avoidance scene recognition device, the vehicle avoidance scene recognition device:

[0242] When a vulnerable traffic participant (VRU) is determined as a target close to the ego vehicle, a driving lane of the vehicle in the historical scene segment is determined;

[0243] Calculating lane deviation data of the vehicle in the driving lane within a preset time period;

[0244] A vehicle avoidance result is determined based on the lane deviation data, wherein the vehicle avoidance result includes whether the vehicle has an avoidance behavior and whether the vehicle has no avoidance behavior.

[0245] Computer program code for performing the operations of the present application may be written in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0246] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0247] The modules described in the embodiments of the present application may be implemented in software or hardware, wherein the name of a module does not necessarily limit the unit itself.

[0248] The computer-readable storage medium provided in this application stores computer-readable program instructions (i.e., a computer program) for executing the aforementioned vehicle avoidance scenario recognition method, thereby resolving the technical issues surrounding vehicle avoidance scenario recognition. Compared to the prior art, the beneficial effects of the computer-readable storage medium provided in this application are similar to those of the vehicle avoidance scenario recognition method provided in the aforementioned embodiments, and are not further elaborated here.

[0249] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned vehicle avoidance scene recognition method when executed by a processor.

[0250] The computer program product provided in this application can solve the technical problem of vehicle avoidance scene recognition. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as the beneficial effects of the vehicle avoidance scene recognition method provided in the above embodiment, and will not be repeated here.

[0251] The above description is only part of the embodiments of the present application and does not limit the patent scope of the present application. All equivalent structural transformations made by using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A vehicle avoidance scene recognition method, characterized in that: The method comprises: When a vulnerable traffic participant (VRU) is determined as a target close to the ego vehicle, a driving lane of the vehicle in the historical scene segment is determined; Calculating lane deviation data of the vehicle in the driving lane within a preset time period; A vehicle avoidance result is determined based on the lane deviation data, wherein the vehicle avoidance result includes whether the vehicle has an avoidance behavior and whether the vehicle has no avoidance behavior.

2. The method according to claim 1, wherein The step of determining the driving lane of the vehicle in the historical scene segment includes: Traversing multiple lane centerlines under the current data frame in the historical scene segment and determining a preliminary associated lane for the vehicle; Calculating the projected distance between the vehicle and the center line of each lane; A driving lane of the vehicle in the historical scene segment is determined based on the projection distance and the preliminary associated lane.

3. The method according to claim 2, wherein The step of determining the driving lane of the vehicle in the historical scene segment based on the projection distance and the preliminary associated lane includes: If the projected distance between the vehicle and the preliminary associated lane is less than or equal to a first preset distance threshold, determining the preliminary associated lane as the driving lane of the vehicle in the historical scene segment; If the projected distance between the vehicle and the initially associated lane is greater than a first preset distance threshold, selecting a minimum projected distance from the projected distances corresponding to the center lines of other lanes; The lane to which the lane centerline corresponding to the minimum projection distance belongs is determined as the driving lane.

4. The method according to claim 1, wherein The step of calculating lane deviation data of the vehicle in the driving lane within a preset time period includes: Determine the distance between the last trajectory point on the current lane centerline and the initial point on the lane centerline; Traversing the data frames between the target frame and the preset frame, calculating the longitudinal displacement and lateral displacement of the current position of the vehicle on the lane centerline in each data frame; If the longitudinal displacement is greater than or equal to the distance value, or less than a second preset distance threshold, the vehicle is determined to be in front of the lane or at the starting position, and no lane offset data is generated; otherwise, the lateral displacement is added to a preset offset track list; Based on the preset deviation trajectory list, lane deviation data of the vehicle in the driving lane is generated.

5. The method according to claim 4, wherein The step of determining a vehicle avoidance result based on the lane deviation data includes: Extracting the first displacement point in the preset offset trajectory list as a reference point, wherein the lateral displacement includes multiple displacement points; Calculate the lateral offset of each displacement point in the preset displacement trajectory list relative to the reference point; Extracting the maximum value of the lateral offsets to obtain the maximum offset value of the lateral displacements; If the maximum offset value is greater than a preset offset threshold, it is determined that the vehicle has an avoidance behavior; otherwise, the vehicle has no avoidance behavior.

6. The method according to claim 1, wherein Before the step of determining that a vulnerable traffic participant VRU is a target close to the vehicle, the method further includes: Obtain historical scene fragments during vehicle driving; Determine a first trajectory point of the VRU and a second trajectory point of the vehicle corresponding to each data frame in the historical scene segment; Based on the first trajectory point and the second trajectory point, it is determined whether the VRU is an object close to the vehicle.

7. The method according to claim 6, wherein The step of determining the first trajectory point of the VRU and the second trajectory point of the vehicle corresponding to each data frame in the historical scene segment includes: Traversing the vehicle indexes and object indexes corresponding to the VRUs in each frame from the target frame to the preset frame, wherein the target frame includes the start frame, and the preset frame is separated from the target frame by a preset frame length; Based on the object index, the vehicle index, and each of the data frames, a plurality of first trajectory points of the VRU and a plurality of second trajectory points of the vehicle within a preset time period are determined.

8. The method according to claim 6, wherein The step of determining whether the VRU is a target close to the vehicle based on the first trajectory point and the second trajectory point includes: calculating a minimum distance between the first trajectory point and the corresponding second trajectory point, and determining displacement data of the VRU relative to the vehicle; When it is determined that the displacement data is within a preset range and the minimum distance is less than a preset distance, the VRU is determined to be a target close to the vehicle.

9. A vehicle avoidance scene recognition device, characterized in that: The device comprises: The first determination module is configured to determine a driving lane of the vehicle in the historical scene segment when a vulnerable traffic participant VRU is determined to be a target close to the vehicle; a calculation module, configured to calculate lane deviation data of the vehicle in the driving lane within a preset time period; The second determination module is configured to determine a vehicle avoidance result based on the lane deviation data, wherein the vehicle avoidance result includes whether the vehicle has an avoidance behavior or not.

10. A vehicle avoidance scene recognition device, characterized in that: The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the vehicle avoidance scene recognition method according to any one of claims 1 to 8.

11. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the vehicle avoidance scene recognition method according to any one of claims 1 to 8 are implemented.

Images