Target determination method, target selection method, domain controller and computer program product

Abstract

The invention provides a target determination method, a target selection method, a domain controller and a computer program product. The method comprises the following steps: firstly, acquiring a detection lane line which is obtained by detecting the surrounding environment of a vehicle and corresponds to a lane of the vehicle and detection frames which respectively correspond to other participating objects in a road, and then for a target detection frame which is intersected with the detection lane line in each detection frame, determining a bending degree corresponding to an intersection position of the target detection frame in the detection lane line; determining the overlapping amount of the target detection frame and the own vehicle lane based on the detection lane line and the target detection frame, determining an overlapping threshold value corresponding to the target detection frame based on the bending degree, and finally judging whether the target detection frame intersected with the detection lane line affects the normal driving of the own vehicle according to the overlapping threshold value and the overlapping amount. And if the overlap amount is greater than an overlap threshold, determining that the target object corresponding to the target detection frame is a self-lane target influencing normal driving of the vehicle, and if the overlap amount is less than or equal to the overlap threshold, determining that the target object corresponding to the target detection frame is an adjacent lane target not influencing normal driving of the vehicle.

Description

Technical Field

[0001] This invention relates to the field of vehicle control technology, and in particular to a target determination method, a target selection method, a domain controller, and a computer program product. Background Technology

[0002] As cars become increasingly intelligent, driver assistance technologies are being widely used. These technologies utilize various sensors installed on the vehicle to sense the surrounding environment while the vehicle is in motion. By collecting data, identifying, detecting, and tracking static and dynamic objects, and combining this data with navigation map data, the system performs calculations and analyses to allow drivers to anticipate potential dangers, effectively increasing driving comfort and safety.

[0003] Cruise control is an important component of driver assistance technology. It is used to continuously scan the road environment ahead using cameras and radar while the vehicle is in motion, select a target vehicle from the vehicles traveling ahead as the target vehicle, and automatically maintain a safe distance from the target vehicle.

[0004] However, while the current cruise control system's target selection method has good accuracy on straight lanes, it is easily affected by vehicles in adjacent lanes when selecting a target on curves due to environmental perception errors or algorithm errors. This can lead to the mistaken selection of vehicles in adjacent lanes as the target, which greatly affects the user's driving experience and driving safety. Summary of the Invention

[0005] Based on this, the present invention provides a target determination method, a target selection method, a domain controller, and a computer program product. By using the target determination method and the target selection method, the accuracy of the vehicle's selection of following targets when driving on curves can be improved, thereby enhancing driving safety and the user's driving experience.

[0006] On one hand, the present invention provides a target determination method, comprising:

[0007] The system obtains the detection lane line corresponding to the vehicle's lane and the detection boxes corresponding to each other participating object in the road, obtained by detecting the environment around the vehicle. For the target detection boxes that intersect with the detection lane line in each detection box, the system determines the degree of curvature corresponding to the position where the detection lane line intersects with the target detection box.

[0008] The overlap between the target detection box and the vehicle lane is determined based on the detected lane line and the target detection box, and the overlap threshold corresponding to the target detection box is determined based on the curvature.

[0009] If the overlap amount is greater than the overlap threshold, then the target object corresponding to the target detection box is determined to be a target in the autonomous lane that affects the normal driving of the vehicle.

[0010] If the overlap amount is less than or equal to the overlap threshold, then the target object corresponding to the target detection box is determined to be an adjacent lane target that does not affect the normal driving of the vehicle.

[0011] Furthermore, in some embodiments, determining the overlap between the target detection box and the vehicle lane based on the detected lane line and the target detection box includes:

[0012] The overlap length between the lower border of the target detection box and the vehicle lane is determined based on the detected lane line and the target detection box;

[0013] The overlap amount is defined as the ratio of the overlap length to the total length of the bottom border of the target detection box.

[0014] Furthermore, in some embodiments, determining the degree of curvature at the intersection point of the detection lane line and the target detection frame includes:

[0015] The intersection point of the detection lane line and the lower border of the target detection box is taken as the intersection position;

[0016] Determine the degree of curvature of the detected lane line at the intersection location.

[0017] Furthermore, in some embodiments, determining the overlap between the target detection box and the vehicle lane based on the detected lane line and the target detection box includes:

[0018] The overlap area between the target detection box and the vehicle lane is determined based on the detected lane line and the target detection box;

[0019] The ratio of the overlapping area to the total area of ​​the target detection box is used as the overlap amount.

[0020] Furthermore, in some embodiments, determining the overlap threshold corresponding to the target detection box based on the degree of curvature includes:

[0021] The overlap threshold corresponding to the target detection box is determined based on the degree of curvature and the vehicle's driving state, wherein the vehicle's driving state includes at least the vehicle's speed.

[0022] The overlap threshold is positively correlated with the degree of curvature, and the overlap threshold is negatively correlated with the vehicle speed.

[0023] Furthermore, in some embodiments, determining the overlap threshold corresponding to the target detection box based on the curvature and the vehicle's driving state, wherein the vehicle's driving state includes at least the vehicle's speed, including:

[0024] The overlap threshold corresponding to the target detection box is determined based on the curvature, the vehicle's driving state, and other detection parameters detected based on the road image. The other detection parameters include, but are not limited to, at least one of the longitudinal distance between the vehicle and the target object and the lateral velocity of the target object.

[0025] The overlap threshold is positively correlated with the longitudinal distance and negatively correlated with the lateral velocity.

[0026] Furthermore, in some embodiments, the degree of bending is expressed by curvature or bending angle.

[0027] On the other hand, the present invention also provides a target selection method, comprising:

[0028] The surrounding environment of the vehicle is detected to obtain the detection lane line corresponding to the vehicle's lane and the detection boxes corresponding to other participating objects in the road.

[0029] Based on the positional relationship between the detection frame and the detection lane line, other participating objects corresponding to all detection frames located in the vehicle lane are identified as vehicle lane targets. The target determination method described above is used to identify vehicle lane targets that may affect the normal driving of the vehicle from among the other participating objects corresponding to each detection frame located in the vehicle lane.

[0030] Select the target that is closest to the longitudinal distance of the vehicle among the targets in the lane.

[0031] Furthermore, in some embodiments, the step of detecting the environment surrounding the vehicle to obtain the detection lane line corresponding to the vehicle's lane and the detection boxes corresponding to other participating objects in the road includes:

[0032] Acquire road point cloud information based on radar and road images based on cameras;

[0033] Based on the road image, lane lines are detected to obtain the detected lane lines corresponding to the vehicle lane;

[0034] Based on the road image and the road point cloud information, correlation detection is performed to obtain the detection boxes corresponding to each other participating object in the road.

[0035] On the other hand, the present invention also provides a domain controller, including: a processor and a memory; wherein the memory stores computer-readable instructions adapted to be loaded by the processor and executed as described in the target determination method and target selection method.

[0036] On the other hand, the present invention also provides a computer program product having at least one instruction stored thereon, wherein the at least one instruction, when executed by a domain controller, implements the steps of the above-described target determination method and target selection method.

[0037] According to the target determination and target selection methods provided by the present invention, when a vehicle is driving on a curve, the detection lane line corresponding to the vehicle's lane and the detection boxes corresponding to other participating objects in the road are obtained by detecting the surrounding environment of the vehicle. Then, for the target detection boxes in each detection box that intersect with the detection lane line, the degree of curvature corresponding to the intersection position of the detection lane line and the target detection box is determined. Then, based on the detection lane line and the target detection box, the overlap between the target detection box and the vehicle's lane is determined, and the overlap threshold corresponding to the target detection box is determined based on the degree of curvature. If the overlap is greater than the overlap threshold, the target detection box is determined to be the target lane. The target is a lane-specific object that affects the normal driving of the vehicle. If the overlap is less than or equal to the overlap threshold, the target object corresponding to the target detection box is determined to be an adjacent lane-specific object that does not affect the normal driving of the vehicle. By setting different overlap thresholds according to the curvature of the vehicle lane, the detection boxes that overlap with the lane lines can be filtered. This method can distinguish between lane-specific objects that will actually affect the normal driving of the vehicle in its own lane and adjacent lane-specific objects that will not affect the normal driving of the vehicle. This avoids the situation where adjacent lane-specific objects are mistakenly selected as the following targets when selecting following targets, thereby improving the accuracy of following target selection, driving safety, and user driving experience.

[0038] It should be understood that the description in the Summary of the Invention is not intended to limit the key or essential features of the embodiments of the present invention, nor is it intended to restrict the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0039] Figure 1 This is a schematic diagram illustrating a scenario of adjacent lane interference provided by an embodiment of the present invention;

[0040] Figure 2 A flowchart illustrating a target determination method provided in an embodiment of the present invention;

[0041] Figure 3 A schematic diagram of an included angle α provided in an embodiment of the present invention;

[0042] Figure 4A flowchart illustrating a target determination method provided in an embodiment of the present invention;

[0043] Figure 5 A flowchart illustrating a target determination method provided in an embodiment of the present invention;

[0044] Figure 6 A flowchart illustrating a target selection method provided in an embodiment of the present invention;

[0045] Figure 7 This is a schematic diagram of the structure of a domain controller provided in an embodiment of the present invention. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0047] In the description of one or more embodiments of the present invention, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc., may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0048] In related technologies, when selecting a following target for cruise control, cameras and radar are typically used to perceive the road environment and detect the vehicle's lane lines and the corresponding bounding boxes of other objects in the road. Then, based on the positional relationship between the bounding boxes of other objects and the vehicle's lane lines, one other object is selected as the following target. Due to environmental perception errors or algorithmic errors, the bounding boxes of other objects detected in adjacent lanes may overlap with the vehicle's lane. The cruise control function may then mistakenly select other objects in adjacent lanes corresponding to bounding boxes that overlap with the vehicle's lane as the following target, leading to incorrect target selection and affecting the user's driving experience and safety. For better understanding, please refer to [link to relevant documentation]. Figure 1 This is a schematic diagram of a scenario involving adjacent lane interference provided by an embodiment of the present invention, such as... Figure 1As shown, the right side is an actual road image, protecting the vehicle 01 in its own lane and other participating objects 02 in adjacent lanes. The left side is a 2D detection result image obtained by the crystal camera and radar perception detection. Curves L1 and L2 are the lane lines corresponding to the vehicle's own lane, and rectangle 03 is the detection box corresponding to other participating objects 02. As shown, the detection box 03 corresponding to other participating objects 02 overlaps with the detected lane line L2. In this way, it is very likely that the detection box 03 will be identified as a target traveling in the vehicle's own lane or a target about to enter the vehicle's own lane. As a result, the cruise control assist function will take the detection box 03 as the following target, leading to the wrong selection of the following target.

[0049] Based on this, the present invention proposes a target determination method and a target selection method. When a vehicle is driving on a curve, it acquires the detection lane line corresponding to the vehicle's lane and the detection boxes corresponding to other participating objects in the road, obtained by detecting the surrounding environment of the vehicle. Then, for the target detection boxes in each detection box that intersect with the detection lane line, it determines the degree of curvature corresponding to the intersection position of the detection lane line with the target detection box. Furthermore, it determines the overlap between the target detection box and the vehicle's lane based on the detection lane line and the target detection box, and determines the overlap threshold corresponding to the target detection box based on the degree of curvature. If the overlap is greater than the overlap threshold, the target detection box is determined to be... The target object is the target in the vehicle's own lane that affects the normal driving of the vehicle. If the overlap is less than or equal to the overlap threshold, the target object corresponding to the target detection box is determined to be an adjacent lane target that does not affect the normal driving of the vehicle. By setting different overlap thresholds according to the curvature of the vehicle's lane, the detection boxes that overlap with the lane lines can be filtered. This method can distinguish the targets in the vehicle's own lane that will actually affect the normal driving of the vehicle in the vehicle's own lane, as well as the adjacent lane targets that will not affect the normal driving of the vehicle. This avoids the situation where adjacent lane targets are mistakenly selected as the following targets when selecting the following targets, thereby improving the accuracy of following target selection, improving driving safety and the user's driving experience.

[0050] Please see Figure 2 , Figure 2 This is a flowchart illustrating a target determination method provided in an embodiment of the present invention. The executing entity of this process can be a program for determining a target in a lane. Alternatively, the executing entity can be a vehicle or domain controller equipped with the aforementioned program, or other devices capable of communicating with the vehicle, domain controller, etc., without specific limitations.

[0051] The following is about Figure 2 The process shown will be described in detail. The target determination method may specifically include the following steps:

[0052] Step S102: Obtain the detection lane line corresponding to the vehicle lane and the detection boxes corresponding to other participating objects in the road obtained by detecting the environment around the vehicle. For the target detection boxes that intersect with the detection lane line in each detection box, determine the degree of curvature corresponding to the position where the detection lane line intersects with the target detection box.

[0053] It should be noted that the target determination method provided in one or more embodiments of the present invention is mainly applied to vehicles operating with following assistance functions. It distinguishes between targets in the vehicle's own lane and targets in adjacent lanes when the vehicle selects a following target based on the following assistance function, thereby avoiding the misselection of targets in adjacent lanes as following targets. The types of following assistance functions can be varied, such as Traffic Jam Assist (TJA), Integrated Cruise Assist (ICA), Adaptive Cruise Control (ACC), etc., and the present invention does not specifically limit these types.

[0054] During the vehicle's operation, based on the surrounding environmental information collected by sensors installed on the vehicle, in this embodiment of the invention, the surrounding environment is detected by the sensors installed on the vehicle to obtain the detection lane line corresponding to the vehicle's lane and the detection boxes corresponding to other participating objects in the road. The detection lane line is real-time lane line information detected by the environment, used to indicate the position of the vehicle's lane line. The detection boxes are other participating objects in the road detected by the environment, used to indicate the positions of other participating objects in the road. These other participating objects can be other vehicles.

[0055] Optionally, the sensors installed on the vehicle can be cameras to capture images of the road surrounding the vehicle. The road images are the images of the road ahead captured by the camera during the operation of the following-assist function. While the vehicle is in motion, the vehicle uses the camera to capture real-time images of the road ahead. Based on these images, it can detect and identify lane lines and detection boxes for other objects on the road. The lane lines are the real-time lane line information identified by the vehicle based on the road images captured by the camera, and the detection boxes are the other objects on the road identified by the vehicle based on the road images captured by the camera.

[0056] In one feasible implementation, a deep-trained lane detection model is used to detect lane lines based on a road image, thereby obtaining detected lane lines; a deep-trained object detection model is used to detect objects based on the road image, thereby obtaining detection boxes corresponding to other participating objects located in the road.

[0057] Furthermore, based on the positional relationship of the detection bounding boxes corresponding to the detected lane line and each other participating object in the two-dimensional coordinate system, the detection bounding box intersecting with the detected lane line is determined as the target detection bounding box, and the degree of curvature corresponding to the intersection of the detected lane line and the target detection bounding box is determined. See also Figure 1 When the vehicle is in a curve, the detection box 03 of other participating objects 02 in the adjacent lane detected by the vehicle 01 overlaps with the lane line of the vehicle. When the detection box intersects with the detection lane line, the existing following target selection algorithm cannot distinguish whether the other participating object corresponding to the detection box is a target traveling in the adjacent lane.

[0058] In one feasible implementation, the degree of curvature at the intersection of the detection lane line and the target detection box can be determined by taking the intersection point of the detection lane line and the lower border of the target detection box as the intersection point and determining the degree of curvature of the detection lane line at the intersection point.

[0059] Optionally, the intersection position can also be the intersection point of the detection lane line and the side border of the target detection box, or it can be the position corresponding to any point of the overlapping part of the detection lane line and the target detection box.

[0060] Step S104: Determine the overlap between the target detection box and the vehicle lane based on the detected lane lines and the target detection box, and determine the overlap threshold corresponding to the target detection box based on the degree of curvature.

[0061] In this embodiment of the invention, after determining the target detection box, the overlap between the target detection box and the vehicle lane is determined based on the positional relationship between the detection lane line and the target detection box, and the overlap threshold corresponding to the target detection box is determined based on the curvature of the intersection position of the detection lane line and the target detection box.

[0062] The overlap amount can refer to the area overlap between the target detection box and the vehicle lane, or the horizontal overlap distance between them. The overlap threshold is a threshold set based on the overlap amount. By setting the overlap threshold, it is possible to determine whether the target object corresponding to the target detection box is located in an adjacent lane based on the overlap amount.

[0063] It is understandable that, such as Figure 1 In curves, due to environmental perception errors or algorithmic errors, other participating objects in adjacent lanes may be detected with overlapping bounding boxes and the vehicle's lane. The more curved the curve, the greater the overlap between the vehicle's lane and the bounding box. By setting an appropriate overlap threshold, bounding boxes that overlap with the vehicle's lane can be distinguished, determining whether the target object corresponding to the bounding box is located in the vehicle's lane or in an adjacent lane.

[0064] In one feasible implementation, a parameter relationship table of curvature degree and overlap threshold can be pre-calibrated. The overlap threshold corresponding to the curvature degree of the curve in different scenarios can be found by looking up the table. The overlap threshold is then used to determine whether the target object corresponding to the target detection box is a target located in an adjacent lane.

[0065] Step S106: If the overlap is greater than the overlap threshold, then the target object corresponding to the target detection box is determined to be a target in the self-lane that affects the normal driving of the self-vehicle.

[0066] Step S108: If the overlap is less than or equal to the overlap threshold, then the target object corresponding to the target detection box is determined to be an adjacent lane target that does not affect the normal driving of the vehicle.

[0067] In this embodiment of the invention, if the overlap between the target detection box and the vehicle lane is greater than the overlap threshold, it indicates that the target object corresponding to the target detection box is likely to be a target traveling in the vehicle lane. In this case, the target object overlapping with the vehicle lane can be determined as a vehicle lane target that can affect the normal driving of the vehicle along the vehicle lane. If the overlap between the target detection box and the vehicle lane is less than or equal to the overlap threshold, it indicates that the target object corresponding to the target detection box is likely to be a target traveling in an adjacent lane. In this case, the target object overlapping with the vehicle lane can be determined as an adjacent lane target that will not affect the normal driving of the vehicle along the vehicle lane.

[0068] In this embodiment of the invention, when the vehicle is traveling on a curve, the detection lane line corresponding to the vehicle's lane and the detection boxes corresponding to other participating objects in the road are obtained by detecting the surrounding environment of the vehicle. Then, for the target detection boxes in each detection box that intersect with the detection lane line, the degree of curvature corresponding to the intersection position of the detection lane line and the target detection box is determined. Then, based on the detection lane line and the target detection box, the overlap between the target detection box and the vehicle's lane is determined, and the overlap threshold corresponding to the target detection box is determined based on the degree of curvature. If the overlap is greater than the overlap threshold, the target object corresponding to the target detection box is determined. For targets in the vehicle's own lane that may affect the vehicle's normal driving, if the overlap is less than or equal to the overlap threshold, the target object corresponding to the target detection box is determined to be an adjacent lane target that does not affect the vehicle's normal driving. By setting different overlap thresholds according to the curvature of the vehicle's lane, detection boxes that overlap with the lane lines can be filtered out. This can distinguish between targets in the vehicle's own lane that may actually affect the vehicle's normal driving in the vehicle's own lane and targets in adjacent lanes that will not affect the vehicle's normal driving. This avoids the situation where adjacent lane targets are mistakenly selected as the following targets when selecting following targets, thereby improving the accuracy of following target selection, enhancing driving safety, and improving the user's driving experience.

[0069] Furthermore, in one or more embodiments of the present invention, the detection lane line includes two single-sided lane lines, one on the left and one on the right, wherein each single-sided lane line is represented by a key point set consisting of multiple lane line key points. Determining the curvature corresponding to the intersection of the detection lane line and the target detection box can specifically be as follows: determining the single-sided lane line intersecting with the target detection box; determining the first key point and the second key point closest to the lower border of the target detection box in the key point set corresponding to the single-sided lane line; determining the third key point and the fourth key point closest to the vehicle in the key point set; determining the first vector pointing from the first key point to the second key point, and the second vector pointing from the third key point to the fourth key point; calculating the angle α between the first vector and the second vector, and using the angle α to represent the curvature corresponding to the intersection of the detection lane line and the target detection box.

[0070] Please see Figure 3 This is a schematic diagram of an included angle α provided in an embodiment of the present invention. Figure 3 As shown, the diagram includes a detection frame 03, a left lane line L1, and a right lane line L2. The left lane line L1 and right lane line L2 each have multiple lane line key points as shown in the diagram. m1 is the first key point closest to the lower border of the detection frame 03, m2 is the second key point closest to the lower border of the detection frame 03, m3 is the third key point closest to the vehicle, and m4 is the fourth key point closest to the vehicle. e2 is a first vector pointing from the first key point to the second key point, and e1 is a second vector pointing from the third key point to the fourth key point. The angle α between the first vector e1 and the second vector e2 is shown in the diagram. In one or more embodiments of the present invention, the angle α can represent the curvature of the detected lane line, where a larger angle α indicates a greater curvature of the detected lane line.

[0071] One feasible implementation method is to determine the overlap between the target detection box and the vehicle lane based on the detected lane line and the target detection box, and determine the overlap threshold corresponding to the target detection box according to the included angle α. If the overlap is greater than the overlap threshold, the target object corresponding to the target detection box is determined to be an interfering target that affects the normal driving of the vehicle. If the overlap is less than or equal to the overlap threshold, the target object corresponding to the target detection box is determined to be an adjacent lane target that does not affect the normal driving of the vehicle.

[0072] In one embodiment, please refer to Figure 4 This is a flowchart illustrating a target determination method provided in an embodiment of the present invention. Figure 4 As shown, it includes the following steps:

[0073] Step S202: Obtain the detection lane line corresponding to the vehicle lane and the detection boxes corresponding to other participating objects in the road obtained by detecting the environment around the vehicle. For the target detection boxes that intersect with the detection lane line in each detection box, determine the degree of curvature corresponding to the position where the detection lane line intersects with the target detection box.

[0074] For step S202, please refer to the detailed description of step S102 in another embodiment of the present invention, which will not be repeated here.

[0075] Step S204: Determine the overlap length between the lower border of the target detection box and the vehicle lane based on the detected lane lines and the target detection box;

[0076] It's easy to understand that if the target detection box is a rectangle, then the bottom border of the target detection box is the base of the rectangle. When one lane line intersects with the target detection box, the portion of the target detection box's bottom border that lies within the vehicle's lane is the overlapping portion, and the length of this portion is the overlap length. For the same target detection box, the longer the overlap length, the greater the probability that the target object corresponding to that detection box is located within the vehicle's lane; conversely, the shorter the overlap length, the lower the probability that the target object corresponding to that detection box is located within the vehicle's lane.

[0077] Step S206: The ratio of the overlap length to the total length of the bottom border of the target detection box is used as the overlap amount;

[0078] In this embodiment of the invention, the ratio of the overlap length to the total length of the bottom border of the target detection box is used as the overlap amount. This overlap amount can represent the ratio of the overlap between the bottom border of the target detection box and the vehicle lane to the bottom border of the target detection box.

[0079] It should be noted that since there are other participating objects of different sizes in the road, the size of the detection boxes corresponding to these other participating objects of different sizes is also different. Therefore, by using the ratio of the overlap length to the total length of the bottom border of the target detection box as the overlap amount, the degree of overlap between the target object corresponding to the target detection box and the vehicle lane can be more accurately represented.

[0080] Step S208: Determine the overlap threshold corresponding to the target detection box based on the degree of curvature and the vehicle's driving state, where the vehicle's driving state includes at least the vehicle's speed.

[0081] In this embodiment of the invention, an overlap threshold is determined based on the curvature of the intersection point of the detection lane line and the target detection box and the vehicle speed. The overlap threshold is positively correlated with the curvature and negatively correlated with the vehicle speed.

[0082] Specifically, the more curved the detection lane line, the larger the corresponding overlap threshold; the more flat the detection lane line, the smaller the corresponding overlap threshold. The faster the vehicle speed, the smaller the corresponding overlap threshold; the faster the vehicle speed, the larger the corresponding overlap threshold.

[0083] The other participating objects corresponding to the target detection box are the target objects.

[0084] Understandably, the greater the curvature of the detection lane line, the more overlapping portions of the detection boxes corresponding to other participating objects in adjacent lanes are within the vehicle's lane. Therefore, a larger overlap threshold is set to accurately distinguish targets in adjacent lanes. Similarly, the smaller the curvature of the detection lane line, the less overlapping portions of the detection boxes corresponding to other participating objects in adjacent lanes are within the vehicle's lane. Therefore, a smaller overlap threshold is set to accurately distinguish targets in adjacent lanes. The faster the vehicle's speed, the shorter the time required for the vehicle to reach the target detection box, and the higher the danger level of the target detection box. Setting a smaller overlap threshold avoids misidentifying targets located in the vehicle's lane as targets located in adjacent lanes, thereby improving safety.

[0085] In one feasible implementation, a suitable overlap threshold fitting equation can be constructed based on the above logic. The overlap threshold is obtained by fitting the degree of curvature and the vehicle speed. The degree of curvature can be represented by curvature or curvature angle.

[0086] Furthermore, an overlap threshold corresponding to the target detection box is determined based on the degree of curvature, the vehicle's driving state, and other detection parameters detected based on the road image. These other detection parameters include, but are not limited to, at least one of the longitudinal distance between the vehicle and the target object and the lateral velocity of the target object. The overlap threshold is negatively correlated with the longitudinal distance and negatively correlated with the lateral velocity.

[0087] Specifically, the smaller the vertical distance, the smaller the corresponding overlap threshold, and the larger the vertical distance, the larger the corresponding overlap threshold; and the smaller the horizontal velocity, the larger the corresponding overlap threshold, and the larger the horizontal velocity, the smaller the corresponding overlap threshold.

[0088] Understandably, the smaller the longitudinal distance, the closer the vehicle and the target object are, the shorter the time required for the vehicle to reach the target detection box, and the higher the danger level of the target detection box. Setting a smaller overlap threshold avoids misidentifying a target object in the vehicle's lane as a target object in an adjacent lane, thus improving safety. Conversely, the smaller the longitudinal distance, the closer the vehicle and the target object are, the shorter the time required for the vehicle to reach the target detection box, and the lower the danger level of the target detection box, so a larger overlap threshold is set. Similarly, the smaller the lateral velocity, the lower the probability that the target object will cut into the vehicle's lane from an adjacent lane, so a larger overlap threshold is set. Conversely, the larger the lateral velocity, the higher the probability that the target object will cut into the vehicle's lane from an adjacent lane, so a larger overlap threshold is set.

[0089] In one feasible implementation, a suitable overlap threshold fitting equation can be constructed based on the above logic. The overlap threshold can be obtained by fitting the degree of curvature, the vehicle speed, the lateral speed corresponding to the target, and the longitudinal distance between the vehicle and the target object. The degree of curvature can be represented by curvature or curvature angle.

[0090] Step S210: If the overlap is greater than the overlap threshold, then the target object corresponding to the target detection box is determined to be a target in the self-lane that affects the normal driving of the self-vehicle.

[0091] Step S212: If the overlap is less than or equal to the overlap threshold, then the target object corresponding to the target detection box is determined to be an adjacent lane target that does not affect the normal driving of the vehicle.

[0092] In this embodiment of the invention, when the vehicle is driving on a curve, the detection lane line corresponding to the vehicle's lane and the detection boxes corresponding to other participating objects in the road are obtained by detecting the surrounding environment of the vehicle. Then, for the target detection boxes in each detection box that intersect with the detection lane line, the degree of curvature corresponding to the intersection position of the detection lane line and the target detection box is determined. Then, based on the detection lane line and the target detection box, the overlap between the target detection box and the vehicle's lane is determined, and based on the degree of curvature and the vehicle's driving state, the overlap threshold corresponding to the target detection box is determined. If the overlap is greater than the overlap threshold, the target detection box is determined to be the target lane. The target is a lane-specific object that affects the normal driving of the vehicle. If the overlap is less than or equal to the overlap threshold, the target object corresponding to the target detection box is determined to be an adjacent lane-specific object that does not affect the normal driving of the vehicle. By setting different overlap thresholds according to the curvature of the vehicle lane, the detection boxes that overlap with the lane lines can be filtered out. This can distinguish between lane-specific objects that will actually affect the normal driving of the vehicle in its own lane and adjacent lane-specific objects that will not affect the normal driving of the vehicle. This avoids the situation where adjacent lane-specific objects are mistakenly selected as the following targets when selecting following targets, thereby improving the accuracy of following target selection, driving safety, and user driving experience.

[0093] In one embodiment, please refer to Figure 5 This is a flowchart illustrating a target determination method provided in an embodiment of the present invention. Figure 5 As shown, it includes the following steps:

[0094] Step S302: Obtain the detection lane line corresponding to the vehicle lane and the detection boxes corresponding to other participating objects in the road obtained by detecting the environment around the vehicle. For the target detection boxes that intersect with the detection lane line in each detection box, determine the degree of curvature corresponding to the position where the detection lane line intersects with the target detection box.

[0095] For step S302, please refer to the detailed description of step S102 in another embodiment of the present invention, which will not be repeated here.

[0096] Step S304: Determine the overlap area between the target detection box and the vehicle lane based on the detected lane lines and the target detection box;

[0097] It's easy to understand that the target detection box is a rectangle. When one side of the lane line intersects with the target detection box, the area of ​​the target detection box within the vehicle's lane is called the overlap area. For the same target detection box, the larger the overlap area, the greater the probability that the target object corresponding to that target detection box is located within the vehicle's lane.

[0098] Step S306: The ratio of the overlapping area to the total area of ​​the target detection box is used as the overlap amount;

[0099] In this embodiment of the invention, the ratio of the overlapping area to the total area of ​​the target detection box is used as the overlap amount, which can represent the ratio of the overlapping area of ​​the target detection box and the vehicle lane to the total area of ​​the target detection box.

[0100] It should be noted that since there are other objects of different sizes in the road, the area of ​​the detection box corresponding to these other objects of different sizes is also different. Therefore, by using the ratio of the overlapping area to the total area of ​​the target detection box as the overlap amount, the degree of overlap between the target object corresponding to the target detection box and the vehicle lane can be more accurately represented.

[0101] Step S308: Determine the overlap threshold corresponding to the target detection box based on the degree of curvature and the vehicle's driving state, where the vehicle's driving state includes at least the vehicle's speed.

[0102] In this embodiment of the invention, an overlap threshold is determined based on the curvature of the intersection point of the detection lane line and the target detection box and the vehicle speed. The overlap threshold is positively correlated with the curvature and negatively correlated with the vehicle speed.

[0103] Specifically, the more curved the detection lane line, the larger the corresponding overlap threshold; the more flat the detection lane line, the smaller the corresponding overlap threshold. The faster the vehicle speed, the smaller the corresponding overlap threshold; the faster the vehicle speed, the larger the corresponding overlap threshold.

[0104] The other participating objects corresponding to the target detection box are the target objects.

[0105] Understandably, the greater the curvature of the detection lane line, the more overlapping portions of the detection boxes corresponding to other participating objects in adjacent lanes are within the vehicle's lane. Therefore, a larger overlap threshold is set to accurately distinguish targets in adjacent lanes. Similarly, the smaller the curvature of the detection lane line, the less overlapping portions of the detection boxes corresponding to other participating objects in adjacent lanes are within the vehicle's lane. Therefore, a smaller overlap threshold is set to accurately distinguish targets in adjacent lanes. The faster the vehicle's speed, the shorter the time required for the vehicle to reach the target detection box, and the higher the danger level of the target detection box. Setting a smaller overlap threshold avoids misidentifying targets located in the vehicle's lane as targets located in adjacent lanes, thereby improving safety.

[0106] In one feasible implementation, a suitable overlap threshold fitting equation can be constructed based on the above logic. The overlap threshold is obtained by fitting the degree of curvature and the vehicle speed. The degree of curvature can be represented by curvature or curvature angle.

[0107] Furthermore, an overlap threshold corresponding to the target detection box is determined based on the degree of curvature, the vehicle's driving state, and other detection parameters detected based on the road image. These other detection parameters include, but are not limited to, at least one of the longitudinal distance between the vehicle and the target object and the lateral velocity of the target object. The overlap threshold is negatively correlated with the longitudinal distance and negatively correlated with the lateral velocity.

[0108] Specifically, the smaller the vertical distance, the smaller the corresponding overlap threshold, and the larger the vertical distance, the larger the corresponding overlap threshold; and the smaller the horizontal velocity, the larger the corresponding overlap threshold, and the larger the horizontal velocity, the smaller the corresponding overlap threshold.

[0109] Understandably, the smaller the longitudinal distance, the closer the vehicle and the target object are, the shorter the time required for the vehicle to reach the target detection box, and the higher the danger level of the target detection box. Setting a smaller overlap threshold avoids misidentifying a target object in the vehicle's lane as a target object in an adjacent lane, thus improving safety. Conversely, the smaller the longitudinal distance, the closer the vehicle and the target object are, the shorter the time required for the vehicle to reach the target detection box, and the lower the danger level of the target detection box, so a larger overlap threshold is set. Similarly, the smaller the lateral velocity, the lower the probability that the target object will cut into the vehicle's lane from an adjacent lane, so a larger overlap threshold is set. Conversely, the larger the lateral velocity, the higher the probability that the target object will cut into the vehicle's lane from an adjacent lane, so a larger overlap threshold is set.

[0110] In one feasible implementation, a suitable overlap threshold fitting equation can be constructed based on the above logic. The overlap threshold can be obtained by fitting the degree of curvature, the vehicle speed, the lateral speed corresponding to the target, and the longitudinal distance between the vehicle and the target object. The degree of curvature can be represented by curvature or curvature angle.

[0111] Optionally, the overlap threshold is a value between 0 and 1.

[0112] Step S310: If the overlap is greater than the overlap threshold, then the target object corresponding to the target detection box is determined to be a target in the self-lane that affects the normal driving of the vehicle.

[0113] Step S312: If the overlap is less than or equal to the overlap threshold, then the target object corresponding to the target detection box is determined to be an adjacent lane target that does not affect the normal driving of the vehicle.

[0114] In this embodiment of the invention, when the vehicle is driving on a curve, the detection lane line corresponding to the vehicle's lane and the detection boxes corresponding to other participating objects in the road are obtained by detecting the surrounding environment of the vehicle. Then, for the target detection boxes in each detection box that intersect with the detection lane line, the degree of curvature corresponding to the position where the detection lane line intersects with the target detection box is determined. Then, based on the detection lane line and the target detection box, the overlap between the target detection box and the vehicle's lane is determined, and based on the degree of curvature and the vehicle's driving state, the overlap threshold corresponding to the target detection box is determined. If the overlap is greater than the overlap threshold, the target object corresponding to the target detection box is determined. For targets in the vehicle's own lane that could affect its normal driving, if the overlap is less than or equal to the overlap threshold, the target object corresponding to the target detection box is determined to be an adjacent lane target that will not affect the vehicle's normal driving. By setting different overlap thresholds according to the curvature of the vehicle's lane, detection boxes that overlap with the lane lines can be filtered out. This method can distinguish between targets in the vehicle's own lane that will actually affect the vehicle's normal driving in its own lane and targets in adjacent lanes that will not affect the vehicle's normal driving. This avoids the situation where adjacent lane targets are mistakenly selected as the following targets when selecting following targets, thereby improving the accuracy of following target selection, enhancing driving safety, and improving the user's driving experience.

[0115] Please see Figure 6 , Figure 6 This is a flowchart illustrating a target selection method provided in an embodiment of the present invention. The executing entity of this process can be a program for target selection. Alternatively, the executing entity can be a vehicle or domain controller equipped with the aforementioned program, or other devices capable of communicating with the vehicle, domain controller, etc., without specific limitations. The following describes... Figure 6 The process shown will be described in detail. The target selection method may specifically include the following steps:

[0116] Step S402: Detect the environment around the vehicle to obtain the detection lane line corresponding to the vehicle's lane and the detection boxes corresponding to other participating objects in the road.

[0117] During the vehicle's operation, based on the surrounding environmental information collected by sensors installed on the vehicle, in this embodiment of the invention, the surrounding environment is detected by the sensors installed on the vehicle to obtain the detection lane line corresponding to the vehicle's lane and the detection boxes corresponding to other participating objects in the road. The detection lane line is real-time lane line information detected by the environment, used to indicate the position of the vehicle's lane line. The detection boxes are other participating objects in the road detected by the environment, used to indicate the positions of other participating objects in the road. These other participating objects can be other vehicles.

[0118] Optionally, the sensors installed on the vehicle can be cameras to capture images of the road surrounding the vehicle. The road images are the images of the road ahead captured by the camera during the operation of the following-assist function. While the vehicle is in motion, the vehicle uses the camera to capture real-time images of the road ahead. Based on these images, it can detect and identify lane lines and detection boxes for other objects on the road. The lane lines are the real-time lane line information identified by the vehicle based on the road images captured by the camera, and the detection boxes are the other objects on the road identified by the vehicle based on the road images captured by the camera.

[0119] In one feasible implementation, based on real-time acquired road images, a two-branch detection model is used. One branch performs lane line detection based on the road images to obtain the detection lane lines corresponding to the vehicle lane, and the other branch performs object detection based on the road images to obtain the detection boxes corresponding to each other participating object in the road.

[0120] In one feasible implementation, a deep-trained lane detection model is used to detect lane lines based on a road image, thereby obtaining detected lane lines; a deep-trained object detection model is used to detect objects based on the road image, thereby obtaining detection boxes corresponding to other participating objects located in the road.

[0121] Optionally, the sensors installed on the vehicle may also include radar and cameras. When the sensors installed on the vehicle include radar and cameras, step S402 may specifically be: acquiring road point cloud information based on radar and acquiring road images based on cameras; performing lane line detection based on the road images to obtain the detected lane lines corresponding to the vehicle's lane; and performing association detection based on the road images and road point cloud information to obtain the detection boxes corresponding to each other participating object in the road.

[0122] It's easy to understand that synchronously collecting environmental information from cameras and radar, and then correlating the collected road images and road point cloud information for detection, helps improve detection accuracy.

[0123] Step S404: Based on the positional relationship between the detection box and the detection lane line, determine the other participating objects corresponding to all detection boxes located in the vehicle lane as candidate targets, and use the target determination method described in the above embodiments to determine the vehicle lane targets that will affect the normal driving of the vehicle from the other participating objects corresponding to each detection box partially located in the vehicle lane.

[0124] Specifically, after detecting each detection frame and detection lane line, the detection lane line is used to indicate the position of the vehicle's lane, and the detection frame is used to indicate the position of other participating objects in the road. Based on the positional relationship between the detection frames and detection lane lines, firstly, all other participating objects corresponding to detection frames located entirely within the vehicle's lane are identified as vehicle lane targets. Then, using the target determination method described in the above embodiment, vehicle lane targets that would affect the normal driving of the vehicle are identified from among the other participating objects corresponding to each detection frame partially located within the vehicle's lane. Here, the detection frames partially located within the vehicle's lane refer to the target detection frames that intersect with the detection lane lines.

[0125] Step S406: Select the target closest to your vehicle in each lane as the selected target.

[0126] It's easy to understand that other participants completely located within the vehicle's lane must be lane-specific targets that affect the vehicle's movement. Similarly, lane-specific targets identified by the target determination method among the other participants corresponding to each detection box partially located within the vehicle's lane are also targets that will affect the vehicle's movement. Specifically, after obtaining the lane-specific targets, the target closest to the vehicle is usually selected.

[0127] In this embodiment of the invention, during vehicle operation, road images are acquired, and detection boxes corresponding to lane lines and other participating objects in the road are detected based on the road images. Then, based on the positional relationship between the detection boxes and the lane lines, all other participating objects corresponding to the detection boxes located in the vehicle's lane are identified as candidate targets. The target determination method described in the above embodiments is then used to identify targets in the vehicle's lane that might affect normal vehicle operation from among the other participating objects corresponding to the detection boxes located in the vehicle's lane. This target determination method can distinguish other participating objects located in adjacent lanes, avoiding the erroneous selection of other participating objects in adjacent lanes as the final target, thus improving target selection accuracy. This target selection method can be applied to following assistance functions such as Traffic Jam Assist (TJA), Integrated Cruise Assist (ICA), and Adaptive Cruise Control (ACC) for selecting following targets, improving the accuracy of following target selection while enhancing driving safety and the user's driving experience.

[0128] In some embodiments, the present invention also provides Figure 7 The diagram shows the structure of a domain controller. Figure 7At the hardware level, the domain controller includes a processor 51, an internal bus 52, a network interface 53, memory 54, and non-volatile memory 55, and may also include other hardware required for business operations. This domain controller can be installed in a vehicle. The processor 51 can read the corresponding computer program from the non-volatile memory 55 into memory and run it to implement the aforementioned target determination and target selection methods. This helps improve the accuracy of target selection when following other vehicles, enhancing driving safety and user experience.

[0129] In some embodiments, the present invention also provides a computer program product that may store at least one instruction, which may be loaded and executed by a domain processor as described in the above embodiments, and the specific execution process can be referred to the specific descriptions in the above embodiments, which will not be repeated here.

[0130] Finally, the various embodiments in this invention are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for the system embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the method embodiments.

[0131] The above description is merely an embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A target determination method, comprising: The system obtains the detection lane line corresponding to the vehicle's lane and the detection boxes corresponding to other participating objects in the road, obtained by detecting the environment around the vehicle. For the target detection boxes that intersect with the detection lane line in each detection box, the system determines the degree of curvature corresponding to the position where the detection lane line intersects with the target detection box. The overlap between the target detection box and the vehicle lane is determined based on the detected lane line and the target detection box, and the overlap threshold corresponding to the target detection box is determined based on the curvature. If the overlap amount is greater than the overlap threshold, then the target object corresponding to the target detection box is determined to be a target in the autonomous lane that affects the normal driving of the vehicle. If the overlap amount is less than or equal to the overlap threshold, then the target object corresponding to the target detection box is determined to be an adjacent lane target that does not affect the normal driving of the vehicle.

2. The method according to claim 1, wherein determining the overlap between the target detection box and the vehicle lane based on the detected lane line and the target detection box comprises: The overlap length between the lower border of the target detection box and the vehicle lane is determined based on the detected lane line and the target detection box; The overlap amount is defined as the ratio of the overlap length to the total length of the bottom border of the target detection box.

3. The method according to claim 2, wherein determining the degree of curvature at the intersection point of the detection lane line and the target detection frame comprises: The intersection point of the detection lane line and the lower border of the target detection box is taken as the intersection position; Determine the degree of curvature of the detected lane line at the intersection location.

4. The method according to claim 1, wherein determining the overlap between the detected lane line and the target detection box and the vehicle lane based on the detected lane line and the target detection box comprises: The overlap area between the target detection box and the vehicle lane is determined based on the detected lane line and the target detection box; The ratio of the overlapping area to the total area of ​​the target detection box is used as the overlap amount.

5. The method according to claim 1, wherein determining the overlap threshold corresponding to the target detection box based on the degree of curvature includes: The overlap threshold corresponding to the target detection box is determined based on the degree of curvature and the vehicle's driving state, wherein the vehicle's driving state includes at least the vehicle's speed. The overlap threshold is positively correlated with the degree of curvature, and the overlap threshold is negatively correlated with the vehicle speed.

6. The method according to claim 5, wherein determining the overlap threshold corresponding to the target detection box based on the curvature and the vehicle's driving state, wherein the vehicle's driving state includes at least the vehicle's speed, includes: The overlap threshold corresponding to the target detection box is determined based on the curvature, the vehicle's driving state, and other detection parameters corresponding to the target object detected based on the road image. The other detection parameters include, but are not limited to, at least one of the longitudinal distance between the vehicle and the target object and the lateral velocity of the target object. The overlap threshold is positively correlated with the longitudinal distance and negatively correlated with the lateral velocity.

7. The method according to any one of claims 1-6, wherein the degree of bending is expressed by curvature or bending angle.

8. A target selection method, comprising: The surrounding environment of the vehicle is detected to obtain the detection lane line corresponding to the vehicle's lane and the detection boxes corresponding to other participating objects in the road. Based on the positional relationship between the detection frame and the detection lane line, other participating objects corresponding to all detection frames located in the vehicle lane are determined as vehicle lane targets. The target determination method described in any one of claims 1-7 is used to determine vehicle lane targets that will affect the normal driving of the vehicle from among the other participating objects corresponding to each detection frame partially located in the vehicle lane. Select the target that is closest to the vehicle among all the targets in the lane.

9. The method according to claim 8, wherein detecting the environment surrounding the vehicle to obtain the detection lane line corresponding to the vehicle's lane and the detection boxes corresponding to other participating objects in the road includes: Acquire road point cloud information based on radar and road images based on cameras; Based on the road image, lane lines are detected to obtain the detected lane lines corresponding to the vehicle lane; Based on the road image and the road point cloud information, correlation detection is performed to obtain the detection boxes corresponding to each other participating object in the road.

10. A domain controller, comprising: A processor and a memory; wherein the memory stores computer-readable instructions adapted to be loaded by the processor and to perform the steps of the method as claimed in any one of claims 1 to 9.

11. A computer program product having at least one instruction stored thereon, wherein the at least one instruction, when executed by a domain controller, implements the steps of the method according to any one of claims 1 to 9.

Images