Target limit cut-in identification control method, system, equipment and medium

Abstract

The invention provides a target limit cut-in identification control method, system and device and a medium, and relates to the technical field of vehicle control, the method comprises the following steps: carrying out identification processing on collected image data in front of a vehicle through a first perception model, and obtaining a region of interest in the image data; performing pixel grid image division on the image data in the region of interest through a preset pixel grid size to obtain a to-be-identified pixel grid image set; performing identification processing on each pixel grid image through a second perception model, and determining a target pixel grid image occupied by the target; according to the position information of the target pixel grid image at different moments within the preset duration, the motion state and the motion trend of the target are determined, and the motion trend comprises a cut-in trend and a non-cut-in trend; and determining and executing a corresponding target control action according to the motion state of the vehicle, the motion state and the motion trend of the target and the distance between the target and the vehicle. The objective of the invention is to improve the recognition accuracy and timeliness of a limit cut-in target.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle control, in particular to a target limit cut-in identification control method, system, device and medium. BACKGROUND

[0002] In the current intelligent driving system, due to the limited FOV (Field of View) of the front camera, when the target at the edge of the FOV quickly cuts into the lane in front of the vehicle, the camera cannot capture the entire target, resulting in inaccurate identification and output of the target or its attributes (lateral distance, lateral speed, heading angle, etc.), which leads to the intelligent driving vehicle being unable to respond quickly or making incorrect responses to such limit cut-in scenarios, posing a safety risk. This is mainly due to the current front vision recognition scheme (recognizing the target through a perception model and assigning a 3D box, and determining the target's motion state and trend by judging the size, orientation, and position change of the 3D box), which cannot provide an accurate 3D box of the target at the edge of the FOV (because the model's 3D box is inaccurate due to incomplete or unclear features of the target), resulting in inaccurate output of the target's attributes, leading to missed braking for the cut-in target in limit cut-in scenarios, and false braking caused by misidentification of the target at the edge of the FOV, thus posing a great risk in limit cut-in scenarios. SUMMARY

[0003] Therefore, the present application provides a target limit cut-in identification control method, system, device and medium, aiming to solve or partially solve the problems in the background art.

[0004] The first aspect of the present application provides a target limit cut-in identification control method, which comprises:

[0005] Collecting image data in front of the vehicle at a preset frequency;

[0006] Identifying and processing the image data through a first perception model to obtain a region of interest in the image data;

[0007] Dividing the image data in the region of interest into pixel grid images according to a preset pixel grid size to obtain a set of pixel grid images to be identified;

[0008] Identifying and processing each pixel grid image in the set of pixel grid images to be identified through a second perception model to determine a target pixel grid image occupied by a target;

[0009] Determining the motion state and trend of the target according to the position information of the target pixel grid image at different times within a preset time period, wherein the motion trend includes a cut-in trend and a non-cut-in trend;

[0010] Based on the vehicle's motion state, the target's motion state and trend, and the distance between the target and the vehicle, the corresponding target control action is determined and executed.

[0011] A second aspect of this application provides a target limit entry identification and control system, the system comprising:

[0012] The data acquisition module is used to acquire image data from the front of the vehicle at a preset frequency;

[0013] The first recognition processing module is used to recognize and process the image data through a first perception model to obtain the region of interest in the image data;

[0014] The image segmentation module is used to segment the image data in the region of interest into pixel grids using a preset pixel grid size, thereby obtaining a set of pixel grid images to be identified.

[0015] The second recognition processing module is used to recognize each pixel grid image in the set of pixel grid images to be recognized through the second perception model, and to determine the target pixel grid image occupied by the target.

[0016] The motion state determination module is used to determine the motion state and motion trend of the target based on the position information of the target pixel grid image at different times within a preset time period. The motion trend includes cutting-in trend and non-cut-in trend.

[0017] The control module is used to determine and execute corresponding target control actions based on the vehicle's motion state, the target's motion state and trend, and the distance between the target and the vehicle.

[0018] A third aspect of this application provides an electronic device, comprising: a processor, a memory, and a computer program stored in the memory and running on the processor, wherein the computer program, when executed by the processor, implements the steps of a target limit cut-in identification and control method as described in the first aspect of this application.

[0019] The fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of a target limit entry identification and control method as described in the first aspect of this application.

[0020] The target limit entry identification and control method provided in this application has the following advantages:

[0021] This application provides a target limit entry recognition and control method. First, image data from in front of the vehicle is acquired at a preset frequency. A first perception model is used to process the image data to obtain the region of interest (ROI). The ROI image data is then divided into pixel grids using a preset pixel grid size to obtain a set of pixel grid images to be identified. A second perception model is used to process each pixel grid image in the set to be identified, determining the target pixel grid image occupied by the target. Based on the position information of the target pixel grid images at different times within a preset time period, the target's motion state and motion trend are determined, including entry and non-entry trends. Finally, based on the vehicle's motion state, the target's motion state and motion trend, and the distance between the target and the vehicle, the corresponding target control action is determined and executed.

[0022] Therefore, the target extreme approach identification and control method provided in this application no longer identifies targets and determines their attributes (lateral distance, lateral velocity, heading angle, etc.) by creating a 3D bounding box that completely encloses the target. Instead, it breaks down the complete image data in front of the vehicle into individual pixel-level image data. By identifying each pixel-level image data separately, it determines whether each pixel-level image data is occupied by the target, thereby determining whether a target has entered the safe driving range in front of the vehicle and determining the target's motion state. This method can identify the target and its running state even when the target only partially enters the image acquisition area, thus improving the accuracy and timeliness of extreme approach target identification, thereby shortening the reaction time of intelligent driving vehicles to extreme approach scenarios and improving the safety of intelligent driving vehicles. Attached Figure Description

[0023] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 A flowchart illustrating a target limit entry identification and control method according to one embodiment of this application;

[0025] Figure 2 This is a schematic diagram illustrating risk region division and pixel grid image division in a target limit entry identification and control method according to an embodiment of this application;

[0026] Figure 3 This is a flowchart illustrating the early warning control process in a target limit entry identification and control method according to an embodiment of this application;

[0027] Figure 4 This is a schematic diagram of a target limit entry identification control system according to one embodiment of this application. Detailed Implementation

[0028] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0029] refer to Figure 1 , Figure 1 This is a flowchart illustrating a target limit entry identification and control method according to one embodiment of this application. Figure 1 As shown, the method includes:

[0030] Step S1: Collect image data from the front of the vehicle at a preset frequency.

[0031] In this embodiment, a forward-facing camera installed in the vehicle collects image data of the area in front of the vehicle at a preset frequency. The forward-facing camera is preferably installed at the top of the windshield and / or on the front grille of the vehicle. The preset frequency can be set according to the actual scenario and is not specifically limited here, such as 30fps, 60fps, etc. The preset frequency can also be dynamically adjusted according to the risk of different driving scenarios. For example, in rainy or snowy weather, the image data collection frequency is increased, such as to 60fps; in sunny urban roads, the image data collection frequency is decreased, such as to 30fps.

[0032] Step S2: The image data is processed by the first perception model to obtain the region of interest in the image data.

[0033] In this embodiment, since the recognition and processing method for each frame of image data is the same, steps S2 to S4 are described using a single frame of image data as an example. First, this application pre-establishes a trained first perception model. This first perception model is used to identify and determine regions of interest (ROIs) in the image data that are suspected of containing targets; that is, to determine which areas in the image data are suspected of containing targets. These areas are called ROIs. The first perception model can be a two-stage detector, a Transformer-type model, etc., and is not specifically limited here. The single frame of image data from the front of the vehicle, acquired in step S1, is input into the first perception model for recognition processing. The first perception model will identify and output the ROI of the image data, where the ROI refers to the area in the image data where a target is suspected to exist.

[0034] Step S3: Divide the image data in the region of interest into pixel grid images using a preset pixel grid size to obtain a set of pixel grid images to be identified.

[0035] In this embodiment, the application pre-sets the size of the image data into individual pixel grid images, which is the preset pixel grid size, and this preset pixel grid size can be a fixed value. Since the method provided by this application requires dividing the image data into a large number of pixel grid images before recognition processing, it requires more computing resources, and the vehicle-side processor is used for many aspects of data processing. Therefore, to ensure that the method provided by this application can be executed normally under any computing resource conditions, the preset pixel grid size can be a dynamically adjustable value. Specifically, the preset pixel grid size is dynamically adjusted according to the remaining computing resources of the vehicle-side processor. The more computing resources the vehicle-side processor has remaining, the smaller the preset pixel grid size can be set, allowing the image data to be divided into a larger number of pixel grid images; conversely, the fewer computing resources the vehicle-side processor has remaining, the larger the preset pixel grid size can be set, allowing the image data to be divided into a smaller number of pixel grid images.

[0036] In this embodiment, based on the region of interest (ROI) in the image data determined in step S2, the image data within the ROI is segmented. Then, based on a preset pixel grid size, the segmented ROI is divided into pixel grid images, that is, the ROI is divided into individual pixel grid images of a preset pixel grid size. All the obtained pixel grid images form a set of pixel grid images to be identified.

[0037] Step S4: The second perception model is used to identify each pixel grid image in the set of pixel grid images to be identified, and the target pixel grid image occupied by the target is determined.

[0038] In this embodiment, a pre-trained second perception model is first established. This second perception model is used to identify whether a pixel grid image is occupied by a target. It should be understood that since a pixel grid image is a very small area, while a target is a large object (such as a car, truck, SUV, etc.), a pixel grid image being occupied by a target means that a small part of the target is located in the pixel grid image. The second perception model can be a high-precision model such as YOLOv8 or Vision Transformer, and is not specifically limited here. The set of pixel grid images to be identified, determined in step S3, is input into the second perception model. The second perception model will perform identification processing on each pixel grid image in the set of pixel grid images to be identified, determining which pixel grid images are occupied by the target. These pixel grid images occupied by the target are the target pixel grid images.

[0039] Step S5: Based on the position information of the target pixel grid image at different times within a preset time period, determine the motion state and motion trend of the target, wherein the motion trend includes cutting-in trend and non-cut-in trend.

[0040] In this embodiment, using the same implementation method, the target pixel grid image can be determined for each image data at different times. Based on the coordinate information of the target pixel grid image in the image coordinate system, the coordinate information of the target pixel grid image in physical space is determined. Then, by analyzing the position information of the target pixel grid image in physical space at different times within a preset time period, the motion state and trend of the target in physical space are determined. The motion state of the target includes at least the target's lateral velocity, lateral acceleration, heading angle, and position information in physical space.

[0041] Step S6: Based on the vehicle's motion state, the target's motion state and trend, and the distance between the target and the vehicle, determine and execute the corresponding target control action.

[0042] In this embodiment, after determining the target's motion state and motion area through step S5, the distance between the target and the vehicle is determined based on the target's position information in physical space during its motion state. By analyzing the vehicle's motion state, the target's motion state and trend, and the distance between the target and the vehicle, corresponding target control actions are determined and executed to avoid potential collision risks.

[0043] In this embodiment, the target limit entry identification and control method provided in this application is used to identify and control limit entry targets, thereby improving the accuracy and timeliness of limit entry target identification. The limit entry target is defined as a target that is horizontally located within any one of the first, second, and third risk zones, and vertically within 5 meters of itself in front of it. The 5m vertical distance is merely a preferred value and can be adjusted to other values ​​through risk scenario calibration.

[0044] This application provides a target limit entry recognition and control method. First, image data from in front of the vehicle is acquired at a preset frequency. A first perception model is used to process the image data to obtain the region of interest (ROI). The ROI image data is then divided into pixel grids using a preset pixel grid size to obtain a set of pixel grid images to be identified. A second perception model is used to process each pixel grid image in the set to be identified, determining the target pixel grid image occupied by the target. Based on the position information of the target pixel grid images at different times within a preset time period, the target's motion state and motion trend are determined, including entry and non-entry trends. Finally, based on the vehicle's motion state, the target's motion state and motion trend, and the distance between the target and the vehicle, the corresponding target control action is determined and executed.

[0045] Therefore, the target extreme approach identification and control method provided in this application no longer identifies targets and determines their attributes (lateral distance, lateral velocity, heading angle, etc.) by creating a 3D bounding box that completely encloses the target. Instead, it breaks down the complete image data in front of the vehicle into individual pixel-level image data. By identifying each pixel-level image data, it determines whether the target occupies each pixel-level image data, thereby determining whether a target has entered the safe driving range in front of the vehicle and determining the target's motion state. This method can identify targets and their running state even when the target only partially enters the image acquisition area, thus improving the accuracy and timeliness of extreme approach target identification, thereby shortening the reaction time of intelligent driving vehicles to extreme approach scenarios and improving the safety of intelligent driving vehicles.

[0046] In conjunction with the above embodiments, in one implementation, this application also provides a method for identifying and controlling target limit entry. In this method for identifying and controlling target limit entry, the method further includes steps S01 to S03:

[0047] Step 01: Based on the vehicle's driving trajectory and the preset risk area division rules, divide the area in front of the vehicle into a preset number of risk areas.

[0048] In this embodiment, the vehicle's driving trajectory is predicted and determined based on the vehicle's driving status. After determining the vehicle's driving trajectory, the area in front of the vehicle is divided into a preset number of risk zones with different risk levels based on the driving trajectory and preset risk zone division rules.

[0049] Step S02: Determine the current position of the target based on the position information of the current target pixel grid image.

[0050] In this embodiment, the coordinates of the target pixel grid image in physical space are determined based on its coordinates in the image coordinate system. Finally, by analyzing the coordinates of the target pixel grid image in physical space, the current position of the target in physical space is determined.

[0051] Step S03: Based on the current location of the target, determine the target risk area where the target is currently located.

[0052] In this embodiment, the target's current location in the physical space is determined based on the coordinate range of each risk area in the currently defined area in front of the vehicle. The risk area where the target is currently located is the target risk area.

[0053] In this application, if the method further includes steps S01 to S03, step S6 may include: determining and executing a corresponding target control action based on the vehicle's motion state, the target's motion state and motion trend, the distance between the target and the vehicle, and the target risk area.

[0054] In this embodiment, where the target limit entry identification and control method provided in this application also determines the target risk area where the target is currently located, an optional implementation of step S6 is: determining the distance between the target and the vehicle based on the target's position information in physical space during its motion state. By analyzing the vehicle's motion state, the target's motion state and trend, the distance between the target and the vehicle, and the target risk area where the target is currently located, corresponding target control actions are determined and executed to avoid potential collision risks.

[0055] In conjunction with the above embodiments, in one implementation, this application also provides a target limit entry identification and control method. In this target limit entry identification and control method, step S01 may include: shifting the vehicle's driving trajectory to the left and right by a first distance to obtain two first target boundary lines, the driving trajectory being determined based on the vehicle's driving state; shifting each of the first target boundary lines to the outside of the vehicle's location by a second distance to obtain two second target boundary lines; shifting each of the second target boundary lines to the outside of the vehicle's location by a third distance to obtain two third target boundary lines; determining the area within the two first target boundary lines as a first risk area requiring emergency avoidance; determining the area between the first and second target boundary lines on the same side as a second risk area requiring normal avoidance; determining the area between the second and third target boundary lines on the same side as a third risk area requiring warning; and determining the area outside the two third target boundary lines as a safe area.

[0056] In this embodiment, as Figure 2 As shown, the determined vehicle trajectory is offset to the left and right by a first distance respectively, resulting in two first target boundary lines, as follows. Figure 2 The image shows two boundary lines pointed to by the number 2. After obtaining the two first target boundary lines, each first target boundary line is offset outward by a second distance from the vehicle's location to obtain two second target boundary lines, as shown. Figure 2 The two boundary lines indicated by the number 3 are shown in the diagram. After obtaining the two second target boundary lines, each second target boundary line is further offset outward by a third distance from the vehicle's location to obtain two third target boundary lines, as shown in the diagram. Figure 2 The diagram shows two boundary lines pointed to by the number 4. The first, second, and third distances can all be set according to the actual application scenario; no specific limitations are made here, and their values ​​will vary. Then, the area within the two first target boundary lines is determined as the first risk zone requiring emergency evacuation, such as... Figure 2 The area indicated by number ① in the diagram is the first risk zone. Furthermore, the area between the first and second target boundary lines on the same side is designated as the second risk zone requiring general avoidance. Figure 2 The two areas indicated by number ② shown are both second-risk zones. For example, the area between the first target boundary line located on the left side of the vehicle and the second target boundary line located on the left side of the vehicle is designated as a second-risk zone requiring general avoidance; simultaneously, the area between the first target boundary line located on the right side of the vehicle and the second target boundary line located on the right side of the vehicle is designated as a second-risk zone requiring general avoidance. Furthermore, the area between the second target boundary line and the third target boundary line on the same side is designated as a third-risk zone requiring warning, such as... Figure 2The number ③ shown indicates that both areas are third-risk zones. For example, the area between the second and third target boundary lines on the left side of the vehicle is designated as a third-risk zone requiring warning; simultaneously, the area between the second and third target boundary lines on the right side of the vehicle is also designated as a third-risk zone requiring warning. Furthermore, the area outside the two third-risk boundary lines is designated as a safe zone, such as... Figure 2 The two areas pointed to by the number ④ shown are both safe zones. Figure 2 The number 1 shown indicates the position of the hood as seen from the front-view camera of the vehicle. Figure 2 The small squares in the positive direction shown are the pixel grids that have been divided into the image. It should be understood that... Figure 2 To facilitate understanding of the scheme, the entire region was divided into pixel grid images for illustration. However, in the actual processing, only the region of interest in the image data was divided and analyzed using pixel grid images. Figure 2 The numbers 5 and 6 shown refer to the same target's position changing at different times. For example, target X was at the position of number 5 at the previous moment, and target X was at the position of number 6 at the next moment.

[0057] In conjunction with the above embodiments, in one implementation, this application also provides a target limit entry identification and control method. In this target limit entry identification and control method, the vehicle's trajectory is shifted to the left and right by a first distance to obtain two first target boundary lines. This includes: shifting the vehicle's trajectory to the left and right by a first distance to obtain two first boundary lines; and correcting the coordinate data points of the first boundary lines based on their longitudinal distance from the vehicle to obtain the first target boundary lines.

[0058] In this embodiment, since the target is farther away from the vehicle, the impact on the vehicle's driving safety will be smaller. Therefore, in order to improve the accuracy of the obtained target boundary line, this application will modify the target boundary line.

[0059] Specifically, the vehicle's trajectory is first offset to the left and right by a first distance, resulting in two first boundary lines. Each first boundary line is composed of a large number of coordinate data points, which are sequentially connected to form the first boundary line. The correction method is the same for each first boundary line; taking one boundary line as an example: based on the longitudinal distance between each coordinate data point of the first boundary line and the vehicle, the coordinate data points of the first boundary line are corrected to obtain the first target boundary line corresponding to that first boundary line. The specific correction method is as follows: based on the longitudinal distance between the coordinate data point and the vehicle, the lateral distance by which the coordinate data point is offset inward is determined, and then the lateral coordinate of the coordinate data point is offset inward by this lateral distance. The greater the longitudinal distance between the coordinate data point and the vehicle (i.e., the farther the coordinate data point is in front of the vehicle), the greater the lateral distance by which the coordinate data point is offset inward. The corresponding first target boundary line is generated based on all the offset coordinate data points.

[0060] In conjunction with the above embodiments, in one implementation, this application also provides a target limit entry identification and control method. This target limit entry identification and control method further includes: determining, based on the target's current position, a boundary line to be corrected and a boundary segment to be corrected within that boundary line, wherein the boundary line to be corrected belongs to the target boundary line; correcting the boundary segment to be corrected within the boundary line to be corrected based on the target's motion state and motion trend; and correcting the risk area associated with the boundary line to be corrected based on the corrected boundary line.

[0061] In this embodiment, in order to obtain a more accurate boundary line that is adapted to the specific target and thus obtain a more accurate risk area division, this application modifies the target boundary line based on the location of the target.

[0062] Specifically, after determining the current position of the target based on the position information of the current target pixel grid image, the method further determines which target boundary line the target is currently approaching based on the target's current position. It should be understood that "approaching" here means that the target is located on the outside, the target boundary line is adjacent to the target, and the target boundary line is located on the inside of the target. The target boundary line that the target is currently approaching is identified as the boundary line to be corrected. Currently, boundary line correction will only be performed on the boundary lines that the target is approaching. Furthermore, the correction will not be performed on the entire boundary line to be corrected, but only on a segment of the boundary line that the target is currently approaching; this segment is the boundary segment to be corrected within the boundary line to be corrected.

[0063] In this embodiment, the method for determining the boundary segment to be corrected is as follows: determine the ordinate of the current position of the target, and starting from the ordinate, determine a segment of the boundary line of a predetermined length starting from the ordinate as the boundary segment to be corrected. For example, if the current position of the target is (x1, y1), determine the coordinate data point of the boundary line to be corrected with y1 as the ordinate. If there is no y1, determine the coordinate data point of the boundary line to be corrected with a ordinate smaller than y1 that is adjacent to y0. Then, starting from the determined coordinate data point, determine a segment of the boundary line of a predetermined length in front of the boundary line to be corrected as the boundary segment to be corrected.

[0064] In this embodiment, after determining the boundary segment to be corrected within the boundary line to be corrected, the boundary segment is corrected according to the target's motion state and trend. The correction strategy is to expand the boundary segment to be corrected outwards when the target's motion state and trend are considered high-risk, thereby increasing the risk area and enabling earlier warnings and corresponding control actions. Specifically, the correction rules are as follows: when the target's motion trend is an intrusion trend, each coordinate data point in the boundary segment to be corrected is expanded outwards by a corresponding distance; when the target's lateral velocity reaches a first value, and / or, the longitudinal velocity reaches a second value, and / or, the lateral acceleration reaches a third value, and / or, the longitudinal acceleration reaches a fourth value, each coordinate data point in the boundary segment to be corrected is expanded outwards by a corresponding distance.

[0065] In this embodiment, after correcting the boundary segment to be corrected in the boundary line to be corrected, the corrected boundary segment to be corrected is sequentially connected to the other boundary segments in the boundary line to be corrected, thereby obtaining the corrected boundary line to be corrected. For example, the boundary segment to be corrected a2 in the boundary line to be corrected divides the entire boundary line to be corrected A into other boundary segments a1, boundary segment to be corrected a2, and other boundary segments a3. After correcting the boundary segment to be corrected a2 to obtain the corrected boundary segment a2', the other boundary segments a1, the corrected boundary segment to be corrected a2', and other boundary segments a3 are sequentially connected to obtain the corrected boundary line to be corrected. Since each risk area is divided based on the target boundary line, after the boundary line to be corrected, each risk area related to the boundary line to be corrected is also corrected, thereby obtaining the current final risk area division result. Finally, based on the final risk area division result and the current position of the target, the target risk area where the target is located is determined. If multiple vehicles (i.e. multiple targets) approach a target boundary line at different locations along that target boundary line, the target boundary line segment corresponding to the position of each of the multiple targets will be corrected.

[0066] In conjunction with the above embodiments, in one implementation, this application also provides a target limit entry identification and control method. In this target limit entry identification and control method, based on the vehicle's motion state, the target's motion state and trend, the distance between the target and the vehicle, and the target risk area, corresponding target control actions are determined and executed. These actions include: determining the target as a safe target when it is currently in a safe area and its motion trend is not an entry trend, and continuously tracking the target's motion state; issuing a warning to the vehicle occupants when the target is currently in a safe area and its motion trend is an entry trend; issuing a warning to the vehicle occupants when the target's current target risk area is a third risk area; determining the collision time based on the vehicle's motion state, the target's motion state and trend, and the distance between the target and the vehicle when the target's current target risk area is a second risk area; executing a corresponding braking control action based on the collision time; and executing an emergency braking control action and / or other collision avoidance control actions based on the vehicle's motion state, the target's motion state and trend, and the distance between the target and the vehicle when the target's current target risk area is a first risk area.

[0067] In this embodiment, as Figure 3 As shown, if the target is currently located in a safe area and its movement trend is not a cutting-in trend, the target is determined to be a safe target, and its movement state is continuously tracked. If the target is currently located in a safe area but its movement trend is a cutting-in trend, a warning is issued to the vehicle occupants. This warning is the lowest level, Level 1, and the target is highlighted on the instrument panel or AR-HUD to alert the user to the target's approach. The system continues to track the target. If the target is currently located in the third risk zone, a warning is issued to the vehicle occupants. If the target is currently located in the second risk zone, the collision time required for a collision is determined based on the vehicle's movement state, the target's movement state and trend, and the distance between the target and the vehicle. Based on this collision time, the vehicle executes the corresponding braking control action. If the target is currently located in the first risk zone, the current risk level is determined to be very high. In this case, based on the vehicle's movement state, the target's movement state and trend, and the distance between the target and the vehicle, emergency braking control actions and / or other collision avoidance control actions, such as steering, will be executed to avoid a collision.

[0068] In conjunction with the above embodiments, in one implementation, this application also provides a target limit entry identification and control method. In this target limit entry identification and control method, when the target's current target risk area is a third risk area, a warning is issued to the vehicle occupants, including: when the target is currently in the third risk area and the target's movement trend is not an entry trend, determining the target as a first-level risk target and issuing a first-level warning to the vehicle occupants; when the target is currently in the third risk area and the target's movement trend is an entry trend, determining the target as a second-level risk target and issuing a second-level warning to the vehicle occupants; when the target is currently in the third risk area, the target's movement trend is an entry trend, and the target's lateral speed exceeds a first threshold, determining the target as a third-level risk target and issuing a third-level warning to the vehicle occupants.

[0069] In this embodiment, as Figure 3 As shown, if the target is currently located in the third risk zone and its movement trend is not an approaching trend, the target is determined to be a level one risk target, and a level one warning is issued to the user in the vehicle. At this time, the target is highlighted on the instrument panel or AR-HUD to remind the user to pay attention to the target approaching, and the system continues to track the target.

[0070] If the target is currently located in the third risk zone and its movement trend is an intrusion trend, the target is determined to be a second-level risk target, and a second-level warning is issued to the occupants of the vehicle. At this time, the instrument panel or AR-HUD displays the obstacle recognition status and highlights or flashes the warning in color, while also providing an audible warning and a steering wheel vibration warning.

[0071] If the target is currently located in the third risk zone, and the target's movement trend is an inward trend, and the target's lateral speed exceeds the first threshold, the target is determined to be a level three risk target, and a level three warning is issued to the vehicle occupants. At this time, based on the level two warning, mild braking and steering within the lane are performed according to the determined target inward speed and collision time.

[0072] Based on the same inventive concept, one embodiment of this application provides a target limit entry identification and control system, such as... Figure 4 As shown, the target limit entry identification and control system 400 includes:

[0073] Data acquisition module 401 is used to acquire image data from the front of the vehicle at a preset frequency;

[0074] The first recognition processing module 402 is used to recognize and process the image data through a first perception model to obtain the region of interest in the image data;

[0075] Image segmentation module 403 is used to segment the image data in the region of interest into pixel grids using a preset pixel grid size to obtain a set of pixel grid images to be identified;

[0076] The second recognition processing module 404 is used to recognize each pixel grid image in the set of pixel grid images to be recognized through the second perception model, and to determine the target pixel grid image occupied by the target.

[0077] The motion state determination module 405 is used to determine the motion state and motion trend of the target based on the position information of the target pixel grid image at different times within a preset time period. The motion trend includes cutting-in trend and non-cut-in trend.

[0078] The control module 406 is used to determine and execute corresponding target control actions based on the vehicle's motion state, the target's motion state and motion trend, and the distance between the target and the vehicle.

[0079] Optionally, the target limit entry recognition control system 400 also includes:

[0080] The risk zone division module is used to divide the area in front of the vehicle into a preset number of risk zones based on the vehicle's driving trajectory and preset risk zone division rules.

[0081] The position determination module is used to determine the current position of the target based on the position information of the current target pixel grid image;

[0082] The target risk area determination module is used to determine the target risk area where the target is currently located based on the target's current location;

[0083] The control module 406 is further configured to determine and execute corresponding target control actions based on the vehicle's motion state, the target's motion state and motion trend, the distance between the target and the vehicle, and the target's risk area.

[0084] Optional risk area delineation module, including:

[0085] The first target boundary line determination module is used to offset the vehicle's driving trajectory to the left and right sides by a first distance to obtain two first target boundary lines. The driving trajectory is determined based on the vehicle's driving state.

[0086] The second target boundary line determination module is used to offset each of the first target boundary lines outward by a second distance from the location of the vehicle to obtain two second target boundary lines.

[0087] The third target boundary line determination module is used to offset each of the second target boundary lines outward by a third distance from the location of the vehicle to obtain two third target boundary lines.

[0088] The first risk area determination module is used to determine the area within the two first target boundary lines as the first risk area requiring emergency avoidance.

[0089] The second risk area determination module is used to determine the area between the first target boundary line and the second target boundary line on the same side as the second risk area that requires ordinary risk avoidance.

[0090] The third risk area determination module is used to determine the area between the second target boundary line and the third target boundary line on the same side as the third risk area that needs to be warned.

[0091] The safe zone determination module is used to determine the area outside the two third target boundary lines as a safe zone.

[0092] Optionally, the first target boundary line determination module includes:

[0093] The first boundary line determination module is used to offset the vehicle's driving trajectory to the left and right by a first distance respectively to obtain two first boundary lines;

[0094] The first target boundary line determination submodule is used to correct the coordinate data points of the first boundary line according to the distance between each coordinate data point of the first boundary line and the vehicle in the longitudinal direction, so as to obtain the first target boundary line.

[0095] Optionally, the target limit entry recognition control system 400 also includes:

[0096] The module for determining the boundary segment to be corrected is used to determine the boundary line to be corrected that the target is close to and the boundary segment to be corrected in the boundary line to be corrected, based on the current position of the target, wherein the boundary line to be corrected belongs to the target boundary line.

[0097] The first correction module is used to correct the boundary segment to be corrected in the boundary line to be corrected according to the motion state and motion trend of the target.

[0098] The second correction module is used to correct the risk areas associated with the boundary line to be corrected based on the corrected boundary line.

[0099] Optionally, the control module 406 includes:

[0100] The first control module is used to determine that the target is a safe target when the target is currently located in a safe area and the target's movement trend is a non-cutting trend, and to continuously track the target's movement state.

[0101] The second control module is used to issue a warning to the user inside the vehicle when the target is currently located in a safe area and the target's movement trend is a cutting trend;

[0102] The third control module is used to issue a warning to the users inside the vehicle when the target is currently located in the third risk area.

[0103] The collision time determination module is used to determine the collision time based on the vehicle's motion state, the target's motion state and trend, and the distance between the target and the vehicle, when the target's current location in the target risk area is the second risk area.

[0104] The fourth control module is used to execute corresponding braking control actions based on the collision time;

[0105] The fifth control module is used to perform emergency braking control actions and / or other collision avoidance control actions when the target's current target risk area is the first risk area, based on the vehicle's motion state, the target's motion state and motion trend, and the distance between the target and the vehicle.

[0106] Optional, a third control module includes:

[0107] The first early warning module is used to determine the target as a level one risk target and issue a level one early warning to the users in the vehicle when the target is currently located in the third risk zone and the target's movement trend is a non-cutting trend.

[0108] The second early warning module is used to determine the target as a secondary risk target and issue a secondary early warning to the vehicle user when the target is currently located in the third risk zone and the target's movement trend is an intrusion trend.

[0109] The third early warning module is used to determine that the target is a level 3 risk target and to issue a level 3 early warning to the vehicle user when the target is currently located in the third risk zone, the target's movement trend is an intrusion trend, and the target's lateral speed exceeds a first threshold.

[0110] Based on the same inventive concept, one embodiment of this application provides an electronic device, including: a processor, a memory, and a computer program stored in the memory and running on the processor. When the computer program is executed by the processor, it implements the steps of the target limit entry identification and control method as described in the first aspect of this application.

[0111] Based on the same inventive concept, one embodiment of this application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of a target limit entry identification and control method as described in the first aspect of this application.

[0112] As the system implementation is basically similar to the method implementation, it is described in a relatively simple way. For relevant details, please refer to the description of the method implementation.

[0113] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of this application are not limited to the described order of actions, because according to the embodiments of this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily necessary for the embodiments of this application.

[0114] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0115] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, embodiments of this application can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of this application can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0116] This application describes embodiments with reference to flowchart illustrations and / or block diagrams of methods, terminal devices (systems), and computer program products according to embodiments of this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0117] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0118] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0119] Although preferred embodiments of the present application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present application.

[0120] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal device that includes said element.

[0121] The above provides a detailed description of the target limit entry identification and control method, system, device, and medium provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for identifying and controlling target limit entry, characterized in that, The method includes: Image data is collected from the front of the vehicle at a preset frequency; The image data is processed by a first perception model to obtain the region of interest in the image data. The image data in the region of interest is divided into pixel grid images by a preset pixel grid size to obtain a set of pixel grid images to be identified. The second perception model is used to identify each pixel grid image in the set of pixel grid images to be identified, and the target pixel grid image occupied by the target is determined. Based on the position information of the target pixel grid image at different times within a preset time period, the motion state and motion trend of the target are determined, and the motion trend includes cutting-in trend and non-cut-in trend; Based on the vehicle's motion state, the target's motion state and trend, and the distance between the target and the vehicle, the corresponding target control action is determined and executed.

2. The target limit entry identification and control method according to claim 1, characterized in that, The method further includes: Based on the vehicle's driving trajectory and the preset risk area division rules, the area in front of the vehicle is divided into a preset number of risk areas; The current position of the target is determined based on the position information of the target pixel grid image. Based on the target's current location, determine the target's current location within the target risk area; The step of determining and executing corresponding target control actions based on the vehicle's motion state, the target's motion state and trend, and the distance between the target and the vehicle includes: Based on the vehicle's motion state, the target's motion state and trend, the distance between the target and the vehicle, and the target's risk area, the corresponding target control action is determined and executed.

3. The target limit entry identification and control method according to claim 2, characterized in that, Based on the vehicle's driving trajectory and preset risk zone division rules, the area in front of the vehicle is divided into a preset number of risk zones, including: The vehicle's trajectory is shifted to the left and right by a first distance to obtain two first target boundary lines, the trajectory being determined based on the vehicle's driving state; Each of the first target boundary lines is offset outward by a second distance from the location of the vehicle to obtain two second target boundary lines; Each of the second target boundary lines is offset outward by a third distance from the location of the vehicle to obtain two third target boundary lines; The area within the two first target boundary lines is designated as the first risk zone requiring emergency evacuation. The area between the first and second target boundary lines on the same side is identified as the second risk zone requiring general risk avoidance. The area between the boundary lines of the second and third targets on the same side is designated as the third risk area requiring early warning. The area outside the two third target boundary lines is defined as a safe zone.

4. The target limit entry identification and control method according to claim 3, characterized in that, The vehicle's trajectory is shifted to the left and right by a first distance respectively, resulting in two first target boundary lines, including: The vehicle's trajectory is shifted to the left and right by a first distance to obtain two first boundary lines; Based on the longitudinal distance between each coordinate data point of the first boundary line and the vehicle, the coordinate data points of the first boundary line are corrected to obtain the first target boundary line.

5. The target limit entry identification and control method according to claim 4, characterized in that, The method further includes: Based on the current position of the target, determine the boundary line to be corrected that the target is close to and the boundary segment to be corrected within the boundary line to be corrected, wherein the boundary line to be corrected belongs to the target boundary line; Based on the motion state and motion trend of the target, the boundary segment to be corrected in the boundary line to be corrected is corrected. Based on the revised boundary line to be corrected, the risk areas associated with the boundary line to be corrected are corrected.

6. The target limit entry identification and control method according to claim 3, characterized in that, Based on the vehicle's motion state, the target's motion state and trend, the distance between the target and the vehicle, and the target's risk area, determine and execute corresponding target control actions, including: If the target is currently located in a safe area and its movement trend is not an intrusion trend, the target is determined to be a safe target, and its movement state is continuously tracked. If the target is currently located in a safe area and the target's movement trend is an approach trend, a warning will be issued to the users inside the vehicle. If the target is currently located in a third-risk area, a warning will be issued to the users inside the vehicle. If the target is currently located in the second risk zone, the collision time is determined based on the vehicle's motion state, the target's motion state and trend, and the distance between the target and the vehicle. Based on the collision time, execute the corresponding braking control action; If the target is currently located in the first risk zone, based on the vehicle's motion state, the target's motion state and trend, and the distance between the target and the vehicle, an emergency braking control action is executed, and / or other collision avoidance control actions are executed.

7. The target limit entry identification and control method according to claim 6, characterized in that, If the target is currently located in a third-risk area, a warning will be issued to the occupants of the vehicle, including: If the target is currently located in the third risk zone and the target's movement trend is a non-cutting trend, the target is determined to be a first-level risk target, and a first-level warning is issued to the users in the vehicle. If the target is currently located in the third risk zone and the target's movement trend is an intrusion trend, the target is determined to be a second-level risk target, and a second-level warning is issued to the users in the vehicle. If the target is currently located in the third risk zone, the target's movement trend is an inward trend, and the target's lateral speed exceeds the first threshold, the target is determined to be a level three risk target, and a level three warning is issued to the users in the vehicle.

8. A target limit entry identification and control system, characterized in that, The system includes: The data acquisition module is used to acquire image data from the front of the vehicle at a preset frequency; The first recognition processing module is used to recognize and process the image data through a first perception model to obtain the region of interest in the image data; The image segmentation module is used to segment the image data in the region of interest into pixel grids using a preset pixel grid size, thereby obtaining a set of pixel grid images to be identified. The second recognition processing module is used to recognize each pixel grid image in the set of pixel grid images to be recognized through the second perception model, and to determine the target pixel grid image occupied by the target. The motion state determination module is used to determine the motion state and motion trend of the target based on the position information of the target pixel grid image at different times within a preset time period. The motion trend includes cutting-in trend and non-cut-in trend. The control module is used to determine and execute corresponding target control actions based on the vehicle's motion state, the target's motion state and trend, and the distance between the target and the vehicle.

9. An electronic device, characterized in that, include: A processor, a memory, and a computer program stored in the memory and running on the processor, wherein the computer program, when executed by the processor, implements the steps of a target limit cut-in identification and control method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the target limit entry identification and control method as described in any one of claims 1 to 7.

Images