Detection Method, Device, Equipment and Storage Medium for Obstacle Cutting-in Behavior

Abstract

The present invention relates to the field of autonomous driving technology, and discloses a method, device, equipment and storage medium for detecting the behavior of an obstacle cutting in, so as to improve the recognition accuracy of the cutting-in behavior. The method for detecting the behavior of an obstacle cutting in includes: obtaining the autonomous driving data to be processed and screening it to obtain at least one candidate obstacle; performing position detection on at least one candidate obstacle based on the position information of the target vehicle; if at least one target obstacle has a suspected cutting-in behavior towards the target vehicle, then performing lane change detection on at least one target obstacle and the target vehicle; if at least one target obstacle meets the object movement screening condition and the target vehicle meets the driving state screening condition, it is determined that at least one target obstacle has a cutting-in behavior towards the target vehicle.

Description

Technical Field

[0001] The present invention relates to the technical field of autonomous driving, and particularly to a method, device, equipment and storage medium for detecting the behavior of an obstacle cutting in. Background Art

[0002] In recent years, the number of motor vehicles in China has increased sharply, and illegal and irregular phenomena in road traffic have also occurred frequently. The behavior of a vehicle cutting in is a common uncivilized driving behavior in people's travel. The behavior of a vehicle cutting in refers to the behavior of a motor vehicle driver overtaking by borrowing a lane or occupying the oncoming lane and cutting in the waiting vehicles when the motor vehicle in front is stopped in a queue or moving slowly.

[0003] In the traditional technology, the discrimination of the behavior of a vehicle cutting in mainly relies on the manual judgment of traffic police or drivers, and this manual judgment method is time-consuming and laborious and has low efficiency. Summary of the Invention

[0004] The present invention provides a method, device, equipment and storage medium for detecting the behavior of an obstacle cutting in, which is used to improve the recognition accuracy of the cutting-in behavior.

[0005] In a first aspect of the present invention, a method for detecting the behavior of an obstacle cutting in is provided, including: obtaining autonomous driving data to be processed, and screening the autonomous driving data to obtain at least one candidate obstacle; performing position detection on the at least one candidate obstacle based on the position information of a target vehicle; if at least one target obstacle among the at least one candidate obstacle has a suspected cutting-in behavior towards the target vehicle, performing lane change detection on the at least one target obstacle and the target vehicle; if the at least one target obstacle meets the object movement screening condition, and the target vehicle meets the driving state screening condition, determining that the at least one target obstacle has a cutting-in behavior towards the target vehicle, where the object movement screening condition is used to indicate that the at least one target obstacle does not have a lane-changing behavior towards the target vehicle, and the driving state screening condition is used to indicate that the target vehicle is in a straight-ahead driving state.

[0006] In a feasible implementation manner, obtaining the to-be-processed autonomous driving data and screening the autonomous driving data to obtain at least one candidate obstacle includes: obtaining the to-be-processed autonomous driving data, where the autonomous driving data includes road video data, the real-time speed of the target vehicle, and radar ranging data, and the road video data includes the target vehicle and at least one initial obstacle; if the average speed corresponding to the real-time speed of the target vehicle is greater than or equal to a preset average speed within a preset detection duration, then based on the road video data, generating the orientation angle difference between the target vehicle and each initial obstacle; based on the radar ranging data, generating the shortest distance between the body contour of the target vehicle and the outer contour of each initial obstacle; determining the initial obstacles that meet the preset screening rules among the at least one initial obstacle as candidate obstacles to obtain at least one candidate obstacle, where the preset screening rules are used to indicate that the orientation angle difference is less than or equal to a preset angle difference and the shortest distance is less than or equal to a preset distance.

[0007] In a feasible implementation manner, the step of, if the average speed corresponding to the real-time speed of the target vehicle is greater than or equal to a preset average speed within a preset detection duration, then based on the road video data, generating the orientation angle difference between the target vehicle and each initial obstacle includes: if the average speed corresponding to the real-time speed of the target vehicle is greater than or equal to a preset average speed within a preset detection duration, then frame the road video data to obtain a set of target road images, where the set of target road images is used to indicate consecutive road video frames, and each target road image in the set of target road images includes the target vehicle and the at least one initial obstacle; perform an aerial view transformation on each target road image to generate the body contour corresponding to the target vehicle and the outer contour corresponding to each initial obstacle in each target road image; determine the vehicle driving direction corresponding to the target vehicle and the object movement direction corresponding to each initial obstacle based on the set of target road images; generate a vehicle orientation line corresponding to the target vehicle based on the center point corresponding to the body contour and the vehicle driving direction, and generate an obstacle orientation line corresponding to each initial obstacle based on the center point corresponding to the outer contour of each initial obstacle and the corresponding object movement direction; obtain the orientation angle difference between the target vehicle and each initial obstacle based on the vehicle orientation line and the obstacle orientation line corresponding to each initial obstacle.

[0008] In a feasible implementation manner, the position detection of the at least one candidate obstacle based on the position information of the target vehicle includes: obtaining the position information corresponding to the at least one candidate obstacle based on the autonomous driving data; generating a lateral distance and a longitudinal distance between the center point of the target vehicle and the center point of each candidate obstacle based on the position information of the target vehicle and the position information corresponding to the at least one candidate obstacle; adding the candidate obstacles that meet the first preset cutting-in condition in the at least one candidate obstacle to the candidate obstacle set, where the first preset cutting-in condition is used to indicate that the lateral distance is greater than or equal to a first preset lane width and less than or equal to a second preset lane width, and the longitudinal distance is greater than or equal to a preset distance behind the vehicle and less than or equal to a first preset distance in front of the vehicle; if there is at least one target obstacle in the candidate obstacle set that meets the second preset cutting-in condition, it is determined that the at least one target obstacle has a suspected cutting-in behavior on the target vehicle, where the second preset cutting-in condition is used to indicate that the lateral distance is less than the first preset lane width within a preset cutting-in time period, and the longitudinal distance is greater than or equal to a preset center distance and less than or equal to a second preset distance in front of the vehicle, and the preset center distance is used to indicate the distance between the center point of the target vehicle and the front bumper.

[0009] In a feasible implementation manner, if at least one target obstacle in the at least one candidate obstacle has a suspected cutting-in behavior on the target vehicle, then lane change detection is performed on the at least one target obstacle and the target vehicle, including: if at least one target obstacle in the at least one candidate obstacle has a suspected cutting-in behavior on the target vehicle, then lane changing behavior detection is performed on the at least one target obstacle to obtain a lane changing behavior detection result; and driving state detection is performed on the target vehicle to obtain a driving state detection result.

[0010] In a feasible implementation manner, the lane changing behavior detection of the at least one target obstacle to obtain a lane changing behavior detection result includes: determining whether there is a change in the first lane corresponding to the target vehicle within a preset time window, where the time window is used to indicate consecutive road video frames; if there is no change in the first lane, determining whether the first lane in the first road video frame of the time window is consistent with the second lane corresponding to the at least one target obstacle; if not, determining whether the first lane in other road video frames of the time window is consistent with the second lane corresponding to the at least one target obstacle; if consistent, determining whether the corresponding relationship between the first lane and the second lane is a lane changing relationship; if not a lane changing relationship, determining that the lane changing behavior detection result is that the at least one target obstacle does not have a lane changing behavior on the target vehicle; if it is a lane changing relationship, determining that the lane changing behavior detection result is that the at least one target obstacle has a lane changing behavior on the target vehicle.

[0011] In a feasible implementation manner, the driving state detection of the target vehicle to obtain a driving state detection result includes: determining whether there is a change in the first lane corresponding to the target vehicle within a preset time window, where the time window is used to indicate consecutive road video frames; if there is no change in the first lane, determining that the driving state detection result is that the target vehicle is in a straight driving state; if there is a change in the first lane, determining that the driving state detection result is that the target vehicle is not in a straight driving state.

[0012] A second aspect of the present invention provides a detection device for an obstacle cutting-in behavior, including: an acquisition and screening module, configured to acquire the autonomous driving data to be processed and screen the autonomous driving data to obtain at least one candidate obstacle; a position detection module, configured to perform position detection on the at least one candidate obstacle based on the position information of the target vehicle; a lane detection module, configured to perform lane change detection on the at least one target obstacle and the target vehicle if at least one target obstacle among the at least one candidate obstacles has a suspected cutting-in behavior with respect to the target vehicle; a cutting-in determination module, configured to determine that the at least one target obstacle has a cutting-in behavior with respect to the target vehicle if the at least one target obstacle meets the object movement screening condition and the target vehicle meets the driving state screening condition, where the object movement screening condition is used to indicate that the at least one target obstacle does not have a lane-changing behavior with respect to the target vehicle, and the driving state screening condition is used to indicate that the target vehicle is in a straight driving state.

[0013] In a feasible implementation manner, the acquisition and screening module includes: an acquisition unit, configured to acquire the autonomous driving data to be processed, where the autonomous driving data includes road video data, the real-time speed of the target vehicle, and radar ranging data, and the road video data includes the target vehicle and at least one initial obstacle; an angle difference generation unit, configured to generate an orientation angle difference between the target vehicle and each initial obstacle based on the road video data if the average speed corresponding to the real-time speed of the target vehicle is greater than or equal to a preset average speed within a preset detection duration; a distance generation unit, configured to generate the shortest distance between the body contour of the target vehicle and the outer contour of each initial obstacle based on the radar ranging data; a determination unit, configured to determine the initial obstacles that meet the preset screening rules among the at least one initial obstacle as candidate obstacles to obtain at least one candidate obstacle, where the preset screening rules are used to indicate that the orientation angle difference is less than or equal to a preset angle difference and the shortest distance is less than or equal to a preset distance.

[0014] In a feasible implementation manner, the angle difference generation unit is specifically configured to: if the average speed corresponding to the real-time speed of the target vehicle is greater than or equal to a preset average speed within a preset detection duration, frame the road video data to obtain a set of target road images, where the set of target road images is used to indicate consecutive road video frames, and each target road image in the set of target road images includes the target vehicle and the at least one initial obstacle; perform a bird's-eye view transformation on each target road image to generate a body contour corresponding to the target vehicle and an external contour corresponding to each initial obstacle in each target road image; determine the vehicle driving direction corresponding to the target vehicle and the object movement direction corresponding to each initial obstacle based on the set of target road images; generate a vehicle orientation line corresponding to the target vehicle based on the center point corresponding to the body contour and the vehicle driving direction, and generate an obstacle orientation line corresponding to each initial obstacle based on the center point corresponding to the external contour of each initial obstacle and the corresponding object movement direction; obtain the orientation angle difference between the target vehicle and each initial obstacle based on the vehicle orientation line and the obstacle orientation line corresponding to each initial obstacle.

[0015] In a feasible implementation manner, the position detection module is specifically configured to: based on the autonomous driving data, obtain the position information corresponding to the at least one candidate obstacle; generate a lateral distance and a longitudinal distance between the center point of the target vehicle and the center point of each candidate obstacle based on the position information of the target vehicle and the position information corresponding to the at least one candidate obstacle; add the candidate obstacles that meet the first preset cutting-in condition among the at least one candidate obstacle to a candidate obstacle set, where the first preset cutting-in condition is used to indicate that the lateral distance is greater than or equal to a first preset lane width and less than or equal to a second preset lane width, and the longitudinal distance is greater than or equal to a preset distance behind the vehicle and less than or equal to a first preset distance in front of the vehicle; if there is at least one target obstacle in the candidate obstacle set that meets the second preset cutting-in condition, determine that the at least one target obstacle has a suspected cutting-in behavior on the target vehicle, where the second preset cutting-in condition is used to indicate that the lateral distance is less than the first preset lane width within a preset cutting-in duration, and the longitudinal distance is greater than or equal to a preset center distance and less than or equal to a second preset distance in front of the vehicle, and the preset center distance is used to indicate the distance between the center point of the target vehicle and the front bumper.

[0016] In a feasible implementation manner, the lane detection module includes: a first detection unit, configured to perform a lane-changing behavior detection on the at least one target obstacle if at least one target obstacle among the at least one candidate obstacle has a suspected cutting-in behavior on the target vehicle, to obtain a lane-changing behavior detection result; a second detection unit, configured to perform a driving state detection on the target vehicle to obtain a driving state detection result.

[0017] In a feasible implementation manner, the first detection unit is specifically configured to: determine whether there is a change in the first lane corresponding to the target vehicle within a preset time window, where the time window is used to indicate consecutive road video frames; if there is no change in the first lane, determine whether the first lane in the first frame of the road video frame of the time window is consistent with the second lane corresponding to the at least one target obstacle; if not, determine whether the first lane in other road video frames of the time window is consistent with the second lane corresponding to the at least one target obstacle; if consistent, determine whether the corresponding relationship between the first lane and the second lane is a lane-changing relationship; if it is not a lane-changing relationship, determine that the lane-changing behavior detection result is that the at least one target obstacle does not have a lane-changing behavior towards the target vehicle; if it is a lane-changing relationship, determine that the lane-changing behavior detection result is that the at least one target obstacle has a lane-changing behavior towards the target vehicle.

[0018] In a feasible implementation manner, the second detection unit is specifically configured to: determine whether there is a change in the first lane corresponding to the target vehicle within a preset time window, where the time window is used to indicate consecutive road video frames; if there is no change in the first lane, determine that the driving state detection result is that the target vehicle is in a straight driving state; if there is a change in the first lane, determine that the driving state detection result is that the target vehicle is not in a straight driving state.

[0019] A third aspect of the present invention provides a detection device for an obstacle cutting-in behavior, including: a memory and at least one processor, where instructions are stored in the memory; the at least one processor calls the instructions in the memory to enable the detection device for the obstacle cutting-in behavior to execute the above-mentioned detection method for the obstacle cutting-in behavior.

[0020] A fourth aspect of the present invention provides a computer-readable storage medium, where instructions are stored in the computer-readable storage medium, and when it runs on a computer, it enables the computer to execute the above-mentioned detection method for the obstacle cutting-in behavior.

[0021] In the technical solution provided by the present invention, the to-be-processed autonomous driving data is obtained, and the autonomous driving data is screened to obtain at least one candidate obstacle; the position of at least one candidate obstacle is detected based on the position information of the target vehicle; if at least one target obstacle among the at least one candidate obstacles has a suspected cutting-in behavior towards the target vehicle, lane change detection is performed on at least one target obstacle and the target vehicle; if at least one target obstacle meets the object movement screening condition and the target vehicle meets the driving state screening condition, it is determined that at least one target obstacle has a cutting-in behavior towards the target vehicle. The object movement screening condition is used to indicate that at least one target obstacle does not have a lane-changing behavior towards the target vehicle, and the driving state screening condition is used to indicate that the target vehicle is in a straight driving state. In the embodiments of the present invention, preliminary screening is performed based on autonomous driving data, and then the position of at least one candidate obstacle is detected. If there is at least one target obstacle with a suspected cutting-in behavior towards the target vehicle, lane change detection is performed on at least one target obstacle and the target vehicle, so as to determine whether at least one target obstacle has a cutting-in behavior towards the target vehicle, improving the recognition accuracy of cutting-in behavior, eliminating various suspected cutting-in scenarios, generating a high-precision obstacle cutting-in scenario dataset, and facilitating the test iteration of the autonomous driving algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a schematic diagram of an embodiment of the method for detecting the cutting-in behavior of an obstacle in an embodiment of the present invention;

[0023] Figure 2 It is a schematic diagram of another embodiment of the method for detecting the cutting-in behavior of an obstacle in an embodiment of the present invention;

[0024] Figure 3 It is a schematic diagram of an embodiment of the device for detecting the cutting-in behavior of an obstacle in an embodiment of the present invention;

[0025] Figure 4 It is a schematic diagram of another embodiment of the device for detecting the cutting-in behavior of an obstacle in an embodiment of the present invention;

[0026] Figure 5 It is a schematic diagram of an embodiment of the device for detecting the cutting-in behavior of an obstacle in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0027] The present invention provides a method, device, equipment and storage medium for detecting the cutting-in behavior of an obstacle, which is used to improve the recognition accuracy of the cutting-in behavior.

[0028] In the description, claims and the above-mentioned drawings of the present invention, terms such as "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects, and do not necessarily describe a specific order or sequence. It should be understood that such data used can be interchanged under appropriate circumstances, so that the embodiments described herein can be implemented in an order different from that shown or described herein. In addition, the term "comprising" or "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not necessarily limit to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0029] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, Artificial Intelligence (AI) is to use a digital computer or a machine controlled by a digital computer to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results in terms of theory, method, technology and application system.

[0030] Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. Artificial intelligence software technologies mainly include several major directions such as computer vision technology, robotics, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0031] For ease of understanding, the specific process of the embodiments of the present invention will be described below. Please refer to Figure 1 , an embodiment of the method for detecting the behavior of an obstacle cutting in line in the embodiments of the present invention includes:

[0032] 101. Acquire the autonomous driving data to be processed, and screen the autonomous driving data to obtain at least one candidate obstacle;

[0033] It can be understood that the execution subject of the present invention can be a detection device for the behavior of an obstacle cutting in line, or a terminal, and specifically, it is not limited here. The embodiments of the present invention will be described by taking the terminal as the execution subject as an example.

[0034] Since there are several obstacles in the autonomous driving data to be processed, such obstacles can be easily filtered out during manual judgment, while when the terminal judges such obstacles, it is necessary to preliminarily screen several obstacles, thereby saving the computing resources for subsequent logical judgment.

[0035] 102. Perform position detection on at least one candidate obstacle based on the position information of the target vehicle;

[0036] The terminal determines the azimuth change of each candidate obstacle relative to the target vehicle within a preset time window. If, in the first frame within the preset time window, the candidate obstacle is located in the left or right direction of the target vehicle, and in any remaining frame within the preset time window, the candidate obstacle appears directly in front of the target vehicle, the terminal determines that there is a suspected cutting-in behavior of the candidate obstacle towards the target vehicle.

[0037] 103. If at least one target obstacle among at least one candidate obstacle has a suspected cutting-in behavior towards the target vehicle, lane change detection is performed on at least one target obstacle and the target vehicle;

[0038] The terminal performs lane change detection on at least one target obstacle to determine whether at least one target obstacle has a lane-changing behavior towards the target vehicle. If at least one target obstacle has a lane-changing behavior, the terminal eliminates the scenario data where at least one target obstacle has a lane-changing behavior. The terminal performs lane change detection on the target vehicle to determine whether the target vehicle is going straight. If the target vehicle is going straight, the terminal retains the scenario data where the target vehicle is going straight, thereby eliminating other suspected cutting-in scenarios.

[0039] 104. If at least one target obstacle meets the object movement screening condition and the target vehicle meets the driving state screening condition, it is determined that at least one target obstacle has a cutting-in behavior towards the target vehicle. The object movement screening condition is used to indicate that at least one target obstacle does not have a lane-changing behavior towards the target vehicle, and the driving state screening condition is used to indicate that the target vehicle is in a straight driving state.

[0040] For example, if at least one target obstacle does not have a lane-changing behavior towards the target vehicle and the target vehicle is in a straight driving state, the terminal determines that at least one target obstacle has a cutting-in behavior towards the target vehicle.

[0041] In the embodiments of the present invention, through preliminary screening based on autonomous driving data and then performing position detection on at least one candidate obstacle, if there is at least one target obstacle with a suspected cutting-in behavior towards the target vehicle, lane change detection is performed on at least one target obstacle and the target vehicle, thereby determining whether at least one target obstacle has a cutting-in behavior towards the target vehicle, improving the recognition accuracy of cutting-in behavior, eliminating various suspected cutting-in scenarios, generating a high-precision obstacle cutting-in scenario dataset, and facilitating the testing and iteration of autonomous driving algorithms.

[0042] Please refer to Figure 2 , another embodiment of the method for detecting obstacle cutting-in behavior in the embodiments of the present invention includes:

[0043] 201. Obtain the autonomous driving data to be processed, and filter the autonomous driving data to obtain at least one candidate obstacle;

[0044] Specifically, 1) The terminal obtains the autonomous driving data to be processed. The autonomous driving data includes road video data, the real-time speed of the target vehicle, and radar ranging data. The road video data includes the target vehicle and at least one initial obstacle; 2) If the average speed corresponding to the real-time speed of the target vehicle is greater than or equal to the preset average speed within the preset detection duration, the terminal generates the orientation angle difference between the target vehicle and each initial obstacle based on the road video data; 3) The terminal generates the shortest distance between the body contour of the target vehicle and the outer contour of each initial obstacle based on the radar ranging data; 4) The terminal determines the initial obstacles that meet the preset screening rules among at least one initial obstacle as candidate obstacles, and obtains at least one candidate obstacle. The preset screening rules are used to indicate that the orientation angle difference is less than or equal to the preset angle difference and the shortest distance is less than or equal to the preset distance.

[0045] Specifically, step 2) If the average speed corresponding to the real-time speed of the target vehicle is greater than or equal to the preset average speed within the preset detection duration, the terminal generates the orientation angle difference between the target vehicle and each initial obstacle based on the road video data, including: (1) If the average speed corresponding to the real-time speed of the target vehicle is greater than or equal to the preset average speed within the preset detection duration, the terminal frames the road video data to obtain a set of target road images, and the set of target road images is used to indicate consecutive road video frames. Each target road image in the set of target road images includes the target vehicle and at least one initial obstacle; (2) The terminal performs an aerial view transformation on each target road image to generate the body contour corresponding to the target vehicle and the outer contour corresponding to each initial obstacle in each target road image; (3) The terminal determines the vehicle driving direction corresponding to the target vehicle and the object movement direction corresponding to each initial obstacle based on the set of target road images; (4) The terminal generates a vehicle orientation line corresponding to the target vehicle based on the center point corresponding to the body contour and the vehicle driving direction, and generates an obstacle orientation line corresponding to each initial obstacle based on the center point corresponding to the outer contour of each initial obstacle and the corresponding object movement direction; (5) The terminal obtains the orientation angle difference between the target vehicle and each initial obstacle based on the vehicle orientation line and the obstacle orientation line corresponding to each initial obstacle.

[0046] For example, the terminal obtains the autonomous driving data to be processed. The autonomous driving data includes road video data, the real-time speed of the target vehicle, and radar ranging data. The road video data includes the target vehicle and at least one initial obstacle. If the average speed corresponding to the real-time speed of the target vehicle is greater than or equal to the preset average speed of 0.5 m / s within 5 seconds, the terminal frames the road video data to obtain a set of target road images. The set of target road images is used to indicate consecutive road video frames. Each target road image in the set of target road images includes the target vehicle and at least one initial obstacle. The terminal performs an aerial view transformation on each target road image to generate the body contour corresponding to the target vehicle and the outer contour corresponding to each initial obstacle in each target road image. The terminal determines the vehicle driving direction corresponding to the target vehicle and the object movement direction corresponding to each initial obstacle based on the set of target road images. The terminal generates a vehicle orientation line corresponding to the target vehicle based on the center point corresponding to the body contour and the vehicle driving direction, and generates an obstacle orientation line corresponding to each initial obstacle based on the center point corresponding to the outer contour of each initial obstacle and the corresponding object movement direction. The terminal obtains the orientation angle difference between the target vehicle and each initial obstacle based on the vehicle orientation line and the obstacle orientation line corresponding to each initial obstacle. The terminal generates the shortest distance between the body contour of the target vehicle and the outer contour of each initial obstacle based on the radar ranging data. The terminal determines the initial obstacles that meet the preset screening rules among the at least one initial obstacle as candidate obstacles, and obtains at least one candidate obstacle. The preset screening rules are used to indicate that the orientation angle difference is less than or equal to the preset angle difference of 115 degrees, and the shortest distance is less than or equal to the preset distance of 20 meters.

[0047] 202. Perform position detection on at least one candidate obstacle based on the position information of the target vehicle;

[0048] Specifically, (1) the terminal obtains the position information corresponding to at least one candidate obstacle based on the autonomous driving data; (2) the terminal generates the lateral distance and the longitudinal distance between the center point of the target vehicle and the center point of each candidate obstacle based on the position information of the target vehicle and the position information corresponding to at least one candidate obstacle; (3) the terminal adds the candidate obstacles that meet the first preset cut-in condition among at least one candidate obstacle to the candidate obstacle set, and the first preset cut-in condition is used to indicate that the lateral distance is greater than or equal to the first preset lane width and less than or equal to the second preset lane width, and the longitudinal distance is greater than or equal to the preset distance behind the vehicle and less than or equal to the first preset distance in front of the vehicle; (4) if there is at least one target obstacle in the candidate obstacle set that meets the second preset cut-in condition, the terminal determines that at least one target obstacle has a suspected cut-in behavior towards the target vehicle, and the second preset cut-in condition is used to indicate that the lateral distance is less than the first preset lane width within the preset cut-in duration, and the longitudinal distance is greater than or equal to the preset center distance and less than or equal to the second preset distance in front of the vehicle, and the preset center distance is used to indicate the distance between the center point of the target vehicle and the front bumper.

[0049] For example, the terminal obtains the position information corresponding to at least one candidate obstacle based on the autonomous driving data; the terminal generates the lateral distance and the longitudinal distance between the center point of the target vehicle and the center point of each candidate obstacle based on the position information of the target vehicle and the position information corresponding to at least one candidate obstacle; the terminal adds the candidate obstacles that meet the first preset cut-in condition among at least one candidate obstacle to the candidate obstacle set, and the first preset cut-in condition is used to indicate that the lateral distance is greater than or equal to the first preset lane width and less than or equal to the second preset lane width, and the longitudinal distance is greater than or equal to the preset distance behind the vehicle by 10 meters and less than or equal to the first preset distance in front of the vehicle by 5 meters, where the first preset lane width is the width of one lane, such as 2 meters, and the second preset lane width is the width of two lanes, such as 4 meters; if there is at least one target obstacle in the candidate obstacle set that meets the second preset cut-in condition, the terminal determines that at least one target obstacle has a suspected cut-in behavior towards the target vehicle, and the second preset cut-in condition is used to indicate that the lateral distance is less than the first preset lane width within 5 seconds, and the longitudinal distance is greater than or equal to the preset center distance and less than or equal to the second preset distance in front of the vehicle by 10 meters, and the preset center distance is used to indicate the distance between the center point of the target vehicle and the front bumper.

[0050] 203. If at least one target obstacle among at least one candidate obstacle has a suspected cut-in behavior towards the target vehicle, then perform a lane change behavior detection on at least one target obstacle to obtain a lane change behavior detection result;

[0051] Specifically, (1) within a preset time window, the terminal determines whether there is a change in the first lane corresponding to the target vehicle, and the time window is used to indicate consecutive road video frames; (2) if there is no change in the first lane, the terminal determines whether the first lane in the first frame of the road video frames of the time window is consistent with the second lane corresponding to at least one target obstacle; (3) if not, the terminal determines whether the first lane in other road video frames of the time window is consistent with the second lane corresponding to at least one target obstacle; (4) if consistent, the terminal determines whether the corresponding relationship between the first lane and the second lane is a lane-changing relationship; (5) if it is not a lane-changing relationship, the terminal determines that the lane-changing behavior detection result is that at least one target obstacle does not have a lane-changing behavior towards the target vehicle; (6) if it is a lane-changing relationship, the terminal determines that the lane-changing behavior detection result is that at least one target obstacle has a lane-changing behavior towards the target vehicle.

[0052] For example, within a preset time window, the terminal determines whether there is a change in the first lane ID corresponding to the target vehicle, and the time window is used to indicate consecutive road video frames; if there is no change in the first lane ID, the terminal determines whether the first lane ID in the first frame of the road video frames of the time window is consistent with the second lane ID corresponding to at least one target obstacle; if not, the terminal determines whether the first lane ID in other road video frames of the time window is consistent with the second lane ID corresponding to at least one target obstacle; if consistent, the terminal determines whether the corresponding relationship between the first lane ID and the second lane ID is a lane-changing relationship; if it is not a lane-changing relationship, the terminal determines that the lane-changing behavior detection result is that at least one target obstacle does not have a lane-changing behavior towards the target vehicle; if it is a lane-changing relationship, the terminal determines that the lane-changing behavior detection result is that at least one target obstacle has a lane-changing behavior towards the target vehicle.

[0053] 204. Perform a driving state detection on the target vehicle to obtain a driving state detection result;

[0054] Specifically, (1) within a preset time window, the terminal determines whether there is a change in the first lane corresponding to the target vehicle, and the time window is used to indicate consecutive road video frames; (2) if there is no change in the first lane, the terminal determines that the driving state detection result is that the target vehicle is in a straight-ahead driving state; (3) if there is a change in the first lane, the terminal determines that the driving state detection result is that the target vehicle is not in a straight-ahead driving state.

[0055] For example, within a preset time window, the terminal determines whether there is a change in the first lane ID corresponding to the target vehicle, and the time window is used to indicate consecutive road video frames; if there is no change in the first lane ID, the terminal determines that the driving state detection result is that the target vehicle is in a straight-ahead driving state; if there is a change in the first lane ID, the terminal determines that the driving state detection result is that the target vehicle is not in a straight-ahead driving state.

[0056] 205. If at least one target obstacle meets the object movement screening condition and the target vehicle meets the driving state screening condition, it is determined that at least one target obstacle has a cutting-in behavior towards the target vehicle. The object movement screening condition is used to indicate that at least one target obstacle has no lane-changing behavior towards the target vehicle, and the driving state screening condition is used to indicate that the target vehicle is in a straight-ahead driving state.

[0057] For example, if at least one target obstacle has no lane-changing behavior towards the target vehicle and the target vehicle is in a straight-ahead driving state, the terminal determines that at least one target obstacle has a cutting-in behavior towards the target vehicle.

[0058] Optionally, steps 203 to 205 can be replaced with the following steps:

[0059] 1. If at least one target obstacle among at least one candidate obstacle has a suspected cutting-in behavior towards the target vehicle, the terminal performs object movement behavior detection on at least one target obstacle to obtain an object movement behavior detection result;

[0060] Specifically, 1) If at least one target obstacle among at least one candidate obstacle has a suspected cutting-in behavior towards the target vehicle, the terminal performs lane-changing behavior detection on at least one target obstacle to obtain a lane-changing behavior detection result;

[0061] Specifically, it is the same as step 203 and will not be elaborated here.

[0062] 2) If at least one target obstacle has no lane-changing behavior towards the target vehicle, the terminal performs special vehicle identification on at least one target obstacle to obtain a special vehicle identification result.

[0063] Specifically, (1) If at least one target obstacle has no lane-changing behavior towards the target vehicle, the terminal obtains the obstacle image and obstacle audio data corresponding to at least one target obstacle based on the autonomous driving data; (2) If the obstacle image corresponding to at least one target obstacle does not match the preset special vehicle image and the corresponding obstacle audio data does not match the preset special vehicle audio data, the terminal determines that the special vehicle identification result is that at least one target obstacle is not a special vehicle; (3) If the obstacle image corresponding to at least one target obstacle matches the preset special vehicle image and the corresponding obstacle audio data matches the preset special vehicle audio data, the terminal determines that the special vehicle identification result is that at least one target obstacle is a special vehicle.

[0064] It can be understood that special vehicles generally have special signs or special vehicle types. Special vehicles refer to vehicles that perform special services and are equipped with special vehicle license plates, sirens, and marker lights. For example, ambulances, fire trucks, police cars, engineering rescue vehicles, military supervision vehicles, etc.

[0065] For example, if at least one target obstacle does not perform a lane-changing behavior on the target vehicle, the terminal obtains the obstacle image and obstacle audio data corresponding to at least one target obstacle based on the autonomous driving data; if the obstacle image corresponding to at least one target obstacle does not match the preset ambulance image and the corresponding obstacle audio data does not match the preset ambulance audio data, the terminal determines that the special vehicle recognition result is that at least one target obstacle is not a special vehicle; if the obstacle image corresponding to at least one target obstacle matches the preset ambulance image and the corresponding obstacle audio data matches the preset ambulance audio data, the terminal determines that the special vehicle recognition result is that at least one target obstacle is a special vehicle.

[0066] 2. The terminal detects the driving state of the target vehicle to obtain a driving state detection result;

[0067] Specifically, it is the same as step 204 and will not be elaborated here.

[0068] 3. If at least one target obstacle meets the object movement screening condition and the target vehicle meets the driving state screening condition, it is determined that at least one target obstacle performs a cutting-in behavior on the target vehicle. The object movement screening condition is used to indicate that at least one target obstacle does not perform a lane-changing behavior on the target vehicle or at least one target obstacle is not a special vehicle, and the driving state screening condition is used to indicate that the target vehicle is in a straight driving state.

[0069] For example, if at least one target obstacle does not perform a lane-changing behavior on the target vehicle or at least one target obstacle is not a special vehicle, and the target vehicle is in a straight driving state, the terminal determines that at least one target obstacle performs a cutting-in behavior on the target vehicle.

[0070] In the embodiment of the present invention, preliminary screening is performed based on autonomous driving data, and then the position of at least one candidate obstacle is detected. If there is at least one target obstacle that performs a suspected cutting-in behavior on the target vehicle, lane change detection is performed on at least one target obstacle and the target vehicle, so as to determine whether at least one target obstacle performs a cutting-in behavior on the target vehicle, improving the recognition accuracy of the cutting-in behavior, eliminating various suspected cutting-in scenarios, generating a high-precision obstacle cutting-in scenario dataset, and facilitating the test iteration of the autonomous driving algorithm.

[0071] The above describes the method for detecting the cut-in behavior of obstacles in the embodiments of the present invention. Next, the detection device for the cut-in behavior of obstacles in the embodiments of the present invention will be described. Please refer to Figure 3 In one embodiment, the detection device for the cut-in behavior of obstacles in the embodiments of the present invention includes:

[0072] An acquisition and screening module 301, configured to acquire the autonomous driving data to be processed and screen the autonomous driving data to obtain at least one candidate obstacle;

[0073] A position detection module 302, configured to perform position detection on at least one candidate obstacle based on the position information of the target vehicle;

[0074] A lane detection module 303, configured to perform lane change detection on at least one target obstacle and the target vehicle if at least one target obstacle among at least one candidate obstacle has a suspected cut-in behavior towards the target vehicle;

[0075] A cut-in determination module 304, configured to determine that at least one target obstacle has a cut-in behavior towards the target vehicle if at least one target obstacle meets the object movement screening condition and the target vehicle meets the driving state screening condition. The object movement screening condition is used to indicate that at least one target obstacle does not have a lane-changing behavior towards the target vehicle, and the driving state screening condition is used to indicate that the target vehicle is in a straight driving state.

[0076] In the embodiments of the present invention, through preliminary screening based on the autonomous driving data, then performing position detection on at least one candidate obstacle, if there is at least one target obstacle with a suspected cut-in behavior towards the target vehicle, performing lane change detection on at least one target obstacle and the target vehicle, so as to determine whether at least one target obstacle has a cut-in behavior towards the target vehicle, improving the recognition accuracy of the cut-in behavior, eliminating various suspected cut-in scenarios, generating a high-precision obstacle cut-in scenario dataset, and facilitating the test iteration of the autonomous driving algorithm.

[0077] Please refer to Figure 4 In another embodiment, the detection device for the cut-in behavior of obstacles in the embodiments of the present invention includes:

[0078] An acquisition and screening module 301, configured to acquire the autonomous driving data to be processed and screen the autonomous driving data to obtain at least one candidate obstacle;

[0079] A position detection module 302, configured to perform position detection on at least one candidate obstacle based on the position information of the target vehicle;

[0080] The lane detection module 303 is configured to perform lane change detection on at least one target obstacle and the target vehicle if at least one target obstacle among at least one candidate obstacle has a suspected cut-in behavior with respect to the target vehicle.

[0081] The cut-in determination module 304 is configured to determine that at least one target obstacle has a cut-in behavior with respect to the target vehicle if at least one target obstacle meets the object movement screening condition and the target vehicle meets the driving state screening condition. The object movement screening condition is used to indicate that at least one target obstacle does not have a lane-changing behavior with respect to the target vehicle, and the driving state screening condition is used to indicate that the target vehicle is in a straight-ahead driving state.

[0082] Optionally, the acquisition and screening module 301 includes:

[0083] An acquisition unit 3011 is configured to acquire the autonomous driving data to be processed. The autonomous driving data includes road video data, the real-time speed of the target vehicle, and radar ranging data. The road video data includes the target vehicle and at least one initial obstacle.

[0084] An angle difference generation unit 3012 is configured to generate the orientation angle difference between the target vehicle and each initial obstacle based on the road video data if the average speed corresponding to the real-time speed of the target vehicle is greater than or equal to the preset average speed within a preset detection duration.

[0085] A distance generation unit 3013 is configured to generate the shortest distance between the body contour of the target vehicle and the outer contour of each initial obstacle based on the radar ranging data.

[0086] A determination unit 3014 is configured to determine the initial obstacles that meet the preset screening rules among at least one initial obstacle as candidate obstacles, and obtain at least one candidate obstacle. The preset screening rules are used to indicate that the orientation angle difference is less than or equal to the preset angle difference and the shortest distance is less than or equal to the preset distance.

[0087] Optionally, the angle difference generation unit 3012 is specifically configured to:

[0088] If the average speed corresponding to the real-time speed of the target vehicle is greater than or equal to the preset average speed within a preset detection duration, frame the road video data to obtain a set of target road images. The set of target road images is used to indicate consecutive road video frames. Each target road image in the set of target road images includes the target vehicle and at least one initial obstacle.

[0089] Perform an aerial view transformation on each target road image to generate the body contour corresponding to the target vehicle and the outer contour corresponding to each initial obstacle in each target road image.

[0090] Determine the vehicle driving direction corresponding to the target vehicle and the object moving direction corresponding to each initial obstacle based on the set of target road images;

[0091] Generate a vehicle orientation line corresponding to the target vehicle based on the center point corresponding to the vehicle body contour and the vehicle driving direction, and generate an obstacle orientation line corresponding to each initial obstacle based on the center point corresponding to the outer contour of each initial obstacle and the corresponding object moving direction;

[0092] Obtain the orientation angle difference between the target vehicle and each initial obstacle based on the vehicle orientation line and the obstacle orientation line corresponding to each initial obstacle.

[0093] Optionally, the position detection module 302 is specifically configured to:

[0094] Obtain the position information corresponding to at least one candidate obstacle based on the autonomous driving data;

[0095] Generate the lateral distance and the longitudinal distance between the center point of the target vehicle and the center point of each candidate obstacle based on the position information of the target vehicle and the position information corresponding to at least one candidate obstacle;

[0096] Add the candidate obstacles that meet the first preset cutting-in condition in at least one candidate obstacle to the candidate obstacle set, and the first preset cutting-in condition is used to indicate that the lateral distance is greater than or equal to the first preset lane width and less than or equal to the second preset lane width, and the longitudinal distance is greater than or equal to the preset distance behind the vehicle and less than or equal to the first preset distance in front of the vehicle;

[0097] If there is at least one target obstacle in the candidate obstacle set that meets the second preset cutting-in condition, it is determined that there is a suspected cutting-in behavior of at least one target obstacle on the target vehicle. The second preset cutting-in condition is used to indicate that the lateral distance is less than the first preset lane width within the preset cutting-in duration, and the longitudinal distance is greater than or equal to the preset center distance and less than or equal to the second preset distance in front of the vehicle. The preset center distance is used to indicate the distance between the center point of the target vehicle and the front bumper.

[0098] Optionally, the lane detection module 303 includes:

[0099] The first detection unit 3031 is configured to perform a lane-changing behavior detection on at least one target obstacle if there is a suspected cutting-in behavior of at least one target obstacle in at least one candidate obstacle on the target vehicle, and obtain a lane-changing behavior detection result;

[0100] The second detection unit 3032 is configured to perform a driving state detection on the target vehicle and obtain a driving state detection result.

[0101] Optionally, the first detection unit 3031 is specifically configured to:

[0102] Determine whether there is a change in the first lane corresponding to the target vehicle within a preset time window, where the time window is used to indicate consecutive road video frames;

[0103] If there is no change in the first lane, determine whether the first lane in the first frame of the road video frame of the time window is consistent with the second lane corresponding to at least one target obstacle;

[0104] If they are inconsistent, determine whether the first lane in other road video frames of the time window is consistent with the second lane corresponding to at least one target obstacle;

[0105] If they are consistent, determine whether the corresponding relationship between the first lane and the second lane is a lane-changing relationship;

[0106] If it is not a lane-changing relationship, determine that the lane-changing behavior detection result is that at least one target obstacle does not have a lane-changing behavior towards the target vehicle;

[0107] If it is a lane-changing relationship, determine that the lane-changing behavior detection result is that at least one target obstacle has a lane-changing behavior towards the target vehicle.

[0108] Optionally, the second detection unit 3032 is specifically configured to:

[0109] Determine whether there is a change in the first lane corresponding to the target vehicle within a preset time window, where the time window is used to indicate consecutive road video frames;

[0110] If there is no change in the first lane, determine that the driving state detection result is that the target vehicle is in a straight driving state;

[0111] If there is a change in the first lane, determine that the driving state detection result is that the target vehicle is not in a straight driving state.

[0112] In the embodiments of the present invention, based on the preliminary screening of autonomous driving data, the position of at least one candidate obstacle is detected. If there is at least one target obstacle that has a suspected cutting-in behavior towards the target vehicle, lane change detection is performed on at least one target obstacle and the target vehicle, so as to determine whether at least one target obstacle has a cutting-in behavior towards the target vehicle, improve the recognition accuracy of the cutting-in behavior, eliminate various suspected cutting-in scenarios, generate a high-precision obstacle cutting-in scenario data set, and facilitate the test iteration of the autonomous driving algorithm.

[0113] Above Figure 3 And Figure 4 The detection device for the obstacle cutting-in behavior in the embodiments of the present invention is described in detail from the perspective of modular functional entities. Next, the detection device for the obstacle cutting-in behavior in the embodiments of the present invention is described in detail from the perspective of hardware processing.

[0114] Figure 5 FIG. 2 is a schematic structural diagram of a detection device for obstacle cutting-in behavior provided by an embodiment of the present invention. The detection device 500 for obstacle cutting-in behavior may vary greatly due to different configurations or performances, and may include one or more processors (central processing units, CPUs) 510 (for example, one or more processors) and a memory 520, and one or more storage media 530 (for example, one or more mass storage devices) for storing application programs 533 or data 532. Among them, the memory 520 and the storage media 530 may be transient storage or persistent storage. The program stored in the storage media 530 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the detection device 500 for obstacle cutting-in behavior. Further, the processor 510 may be configured to communicate with the storage media 530 and execute a series of instruction operations in the storage media 530 on the detection device 500 for obstacle cutting-in behavior.

[0115] The detection device 500 for obstacle cutting-in behavior may further include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input / output interfaces 560, and / or one or more operating systems 531, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, and so on. Those skilled in the art can understand that Figure 5 the shown structural diagram of the detection device for obstacle cutting-in behavior does not constitute a limitation on the detection device for obstacle cutting-in behavior, and may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0116] The present invention also provides a detection device for obstacle cutting-in behavior. The computer device includes a memory and a processor. When the computer-readable instructions stored in the memory are executed by the processor, the processor executes the steps of the detection method for obstacle cutting-in behavior in the above embodiments.

[0117] The present invention also provides a computer-readable storage medium. The computer-readable storage medium may be a non-volatile computer-readable storage medium, or may also be a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are run on a computer, the computer executes the steps of the detection method for obstacle cutting-in behavior.

[0118] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be described herein again.

[0119] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The foregoing storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs.

[0120] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. These modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for detecting the behavior of an obstacle cutting in line, characterized in that The detection method for the behavior of an obstacle cutting in includes: Obtain the autonomous driving data to be processed, and screen the autonomous driving data to obtain at least one candidate obstacle; Perform position detection on the at least one candidate obstacle based on the position information of the target vehicle; If at least one target obstacle among the at least one candidate obstacle has a suspected cutting-in behavior towards the target vehicle, then perform lane change detection on the at least one target obstacle and the target vehicle; If the at least one target obstacle meets the object movement screening condition, and the target vehicle meets the driving state screening condition, it is determined that the at least one target obstacle has a cutting-in behavior towards the target vehicle. The object movement screening condition is used to indicate that the at least one target obstacle does not have a lane-changing behavior towards the target vehicle, and the driving state screening condition is used to indicate that the target vehicle is in a straight driving state; The performing position detection on the at least one candidate obstacle based on the position information of the target vehicle includes: based on the autonomous driving data, obtain the position information corresponding to the at least one candidate obstacle; based on the position information of the target vehicle and the position information corresponding to the at least one candidate obstacle, generate the lateral distance and the longitudinal distance between the center point of the target vehicle and the center point of each candidate obstacle; add the candidate obstacles that meet the first preset cutting-in condition among the at least one candidate obstacle to the candidate obstacle set. The first preset cutting-in condition is used to indicate that the lateral distance is greater than or equal to the first preset lane width and less than or equal to the second preset lane width, and the longitudinal distance is greater than or equal to the preset distance behind the vehicle and less than or equal to the first preset distance in front of the vehicle; if there is at least one target obstacle in the candidate obstacle set that meets the second preset cutting-in condition, it is determined that the at least one target obstacle has a suspected cutting-in behavior towards the target vehicle. The second preset cutting-in condition is used to indicate that within the preset cutting-in duration, the lateral distance is less than the first preset lane width, and the longitudinal distance is greater than or equal to the preset center distance and less than or equal to the second preset distance in front of the vehicle. The preset center distance is used to indicate the distance between the center point of the target vehicle and the front bumper.

2. The detection method of the obstacle cutting-in behavior according to claim 1, characterized in that, The obtaining the autonomous driving data to be processed, and screening the autonomous driving data to obtain at least one candidate obstacle includes: Obtain the autonomous driving data to be processed. The autonomous driving data includes road video data, the real-time speed of the target vehicle, and radar ranging data. The road video data includes the target vehicle and at least one initial obstacle; If the average speed corresponding to the real-time speed of the target vehicle is greater than or equal to the preset average speed within the preset detection duration, then based on the road video data, generate the orientation angle difference between the target vehicle and each initial obstacle; Based on the radar ranging data, generate the shortest distance between the body contour of the target vehicle and the external contour of each initial obstacle; Determine the initial obstacles that meet the preset screening rules among the at least one initial obstacle as candidate obstacles, obtaining at least one candidate obstacle, where the preset screening rules are used to indicate that the difference in the orientation angles is less than or equal to a preset angle difference and the shortest distance is less than or equal to a preset distance.

3. The detection method of the obstacle cutting-in behavior according to claim 2, characterized in that, If the average speed corresponding to the real-time speed of the target vehicle is greater than or equal to the preset average speed within the preset detection duration, then based on the road video data, generate the difference in the orientation angles between the target vehicle and each initial obstacle, including: If the average speed corresponding to the real-time speed of the target vehicle is greater than or equal to the preset average speed within the preset detection duration, then frame the road video data to obtain a set of target road images, where the set of target road images is used to indicate consecutive road video frames, and each target road image in the set of target road images includes the target vehicle and the at least one initial obstacle; Perform an aerial view transformation on each target road image to generate the body contour corresponding to the target vehicle and the outer contour corresponding to each initial obstacle in each target road image; Based on the set of target road images, determine the vehicle driving direction corresponding to the target vehicle and the object movement direction corresponding to each initial obstacle; Based on the center point corresponding to the body contour and the vehicle driving direction, generate the vehicle orientation line corresponding to the target vehicle, and based on the center point corresponding to the outer contour of each initial obstacle and the corresponding object movement direction, generate the obstacle orientation line corresponding to each initial obstacle; Based on the vehicle orientation line and the obstacle orientation line corresponding to each initial obstacle, obtain the difference in the orientation angles between the target vehicle and each initial obstacle.

4. The detection method for the obstacle cutting-in behavior according to any one of claims 1-3, characterized in that If at least one target obstacle among the at least one candidate obstacle has a suspected cutting-in behavior towards the target vehicle, then perform a lane change detection on the at least one target obstacle and the target vehicle, including: If at least one target obstacle among the at least one candidate obstacle has a suspected cutting-in behavior towards the target vehicle, then perform a lane changing behavior detection on the at least one target obstacle to obtain a lane changing behavior detection result; Perform a driving state detection on the target vehicle to obtain a driving state detection result.

5. The detection method of the obstacle cutting-in behavior according to claim 4, wherein The performing a lane changing behavior detection on the at least one target obstacle to obtain a lane changing behavior detection result includes: Judge whether there is a change in the first lane corresponding to the target vehicle within a preset time window, where the time window is used to indicate consecutive road video frames; If there is no change in the first lane, then judge whether the first lane in the first frame of the road video frame of the time window is consistent with the second lane corresponding to the at least one target obstacle; If they are not consistent, then judge whether the first lane in other road video frames of the time window is consistent with the second lane corresponding to the at least one target obstacle; If they are consistent, then judge whether the corresponding relationship between the first lane and the second lane is a lane changing relationship; If it is not a lane-changing relationship, it is determined that the lane-changing behavior detection result is that there is no lane-changing behavior of the at least one target obstacle with respect to the target vehicle; If it is a lane-changing relationship, it is determined that the lane-changing behavior detection result is that the at least one target obstacle has a lane-changing behavior with respect to the target vehicle.

6. The detection method of the obstacle cutting-in behavior according to claim 4, wherein, The driving state detection of the target vehicle to obtain a driving state detection result includes: Judging whether there is a change in the first lane corresponding to the target vehicle within a preset time window, where the time window is used to indicate consecutive road video frames; If there is no change in the first lane, it is determined that the driving state detection result is that the target vehicle is in a straight driving state; If there is a change in the first lane, it is determined that the driving state detection result is that the target vehicle is not in a straight driving state.

7. A detection device for the behavior of an obstacle cutting in line, characterized in that, The detection device for the obstacle cutting-in behavior includes: An acquisition and screening module, configured to acquire the autonomous driving data to be processed and screen the autonomous driving data to obtain at least one candidate obstacle; A position detection module, configured to perform position detection on the at least one candidate obstacle based on the position information of the target vehicle; A lane detection module, configured to perform lane change detection on the at least one target obstacle and the target vehicle if at least one of the at least one candidate obstacles has a suspected cutting-in behavior with respect to the target vehicle; A cutting-in determination module, configured to determine that the at least one target obstacle has a cutting-in behavior with respect to the target vehicle if the at least one target obstacle meets the object movement screening condition and the target vehicle meets the driving state screening condition, where the object movement screening condition is used to indicate that there is no lane-changing behavior of the at least one target obstacle with respect to the target vehicle, and the driving state screening condition is used to indicate that the target vehicle is in a straight driving state; The position detection module is specifically configured to: based on the autonomous driving data, obtain the position information corresponding to the at least one candidate obstacle; based on the position information of the target vehicle and the position information corresponding to the at least one candidate obstacle, generate the lateral distance and the longitudinal distance between the center point of the target vehicle and the center point of each candidate obstacle; add the candidate obstacles that meet the first preset cutting-in condition in the at least one candidate obstacle to the candidate obstacle set, where the first preset cutting-in condition is used to indicate that the lateral distance is greater than or equal to the first preset lane width and less than or equal to the second preset lane width, and the longitudinal distance is greater than or equal to the preset distance behind the vehicle and less than or equal to the first preset distance in front of the vehicle; if there is at least one target obstacle in the candidate obstacle set that meets the second preset cutting-in condition, it is determined that the at least one target obstacle has a suspected cutting-in behavior with respect to the target vehicle, where the second preset cutting-in condition is used to indicate that the lateral distance is less than the first preset lane width within the preset cutting-in duration, and the longitudinal distance is greater than or equal to the preset center distance and less than or equal to the second preset distance in front of the vehicle, and the preset center distance is used to indicate the distance between the center point of the target vehicle and the front bumper.

8. A detection device for the behavior of an obstacle cutting in line, characterized in that, The detection device for the behavior of an obstacle cutting in line includes: a memory and at least one processor, and instructions are stored in the memory; The at least one processor calls the instructions in the memory to enable the detection device for the behavior of an obstacle cutting in line to execute the detection method for the behavior of an obstacle cutting in line according to any one of claims 1-6.

9. A computer-readable storage medium having instructions stored thereon, characterized in that, When the instructions are executed by the processor, the detection method for the behavior of an obstacle cutting in line according to any one of claims 1-6 is implemented.

Images