A lane line tracking method, apparatus, terminal device and storage medium

Abstract

This invention discloses a lane line tracking method, apparatus, terminal device, and storage medium. The method includes: acquiring a current set of lane lines and a set of historical lane lines corresponding to each current lane line; performing a lane line confidence subset acquisition action on each historical lane line set to obtain a lane line confidence set; and obtaining a lane line tracking result by combining the current set of lane lines with the historical set of lane lines based on the lane line confidence set. The lane line confidence subset acquisition action includes: for any current lane line, acquiring a set of distances between matching sampling points and a set of angles between matching sampling points for each historical lane line, based on the current lane line and each historical lane line; obtaining a lane line distance confidence based on the set of distances between matching sampling points; obtaining a lane line angle confidence based on the set of angles between matching sampling points; and obtaining a lane line confidence subset based on the lane line distance confidence and the lane line angle confidence to improve the accuracy of lane line tracking.

Description

Technical Field

[0001] This invention relates to the field of lane tracking, and more particularly to a lane tracking method, apparatus, terminal device, and storage medium. Background Technology

[0002] In the field of autonomous driving, the real-time construction and updating of high-precision maps is a crucial foundation for achieving accurate vehicle positioning and decision-making. Since its release, the MapTR framework has played an increasingly important role in autonomous driving after extensive application and research. MapTR's core function in autonomous driving is to transform the perceived physical environment into a machine-understandable, structured "language"—a vectorized high-precision map—in real time. This provides crucial static environmental information for vehicle navigation and decision-making. MapTR not only unifies complex map elements (lane lines, pedestrian crossings, etc.) into "point sets," greatly simplifying model learning and improving accuracy, but its lightweight version also maintains high accuracy while boasting significantly faster inference speeds than previous solutions, achieving high-quality real-time mapping. The generated vectorized maps directly support downstream path planning and decision-making. Due to its various advantages, MapTR has been adopted by numerous intelligent driving algorithm companies in their mass-production algorithms.

[0003] However, MapTR still has some problems in practical engineering applications: MapTR senses lane lines frame by frame, and there is no correlation between each frame, so the lane lines will jump and jitter, and the accuracy of sensing lane lines at a distance will decrease, making it easy to falsely detect or lose them. In complex scenarios, there are even more false detections, which in turn leads to the loss of lane line tracking. Summary of the Invention

[0004] This invention provides a lane line tracking method, apparatus, terminal device, and storage medium, which can solve the above-mentioned problems and improve the accuracy of lane line tracking.

[0005] This invention provides a lane line tracking method, comprising: Obtain the current lane line set of the road segment to be tracked and the historical lane line set corresponding to each current lane line in the current lane line set; For each of the historical lane line sets, a lane line confidence subset acquisition action is performed to obtain a lane line confidence subset for each of the current lane lines. Then, based on all the lane line confidence subsets, a lane line confidence set corresponding to the current lane line set is obtained. Based on the lane line confidence set, the current lane line set is matched with the historical lane line set to obtain the lane line tracking result of the road segment to be tracked; The lane line confidence subset acquisition action includes: For any current lane line, based on the current lane line and each historical lane line in the corresponding historical lane line set, obtain the set of distances between matching sampling points and the set of included angles between matching sampling points for each historical lane line; The lane line distance confidence level of each historical lane line is obtained by using the distance set of the matching sampling points corresponding to each historical lane line and the first preset lane line distance threshold. The confidence level of the lane line angle of each historical lane line is obtained based on the set of angles of the matching sampling points corresponding to each historical lane line and the preset lane line angle threshold. Based on the lane distance confidence and lane angle confidence of each historical lane line, the lane line confidence of each historical lane line is obtained, so as to obtain a subset of the lane line confidence of the current lane line according to all the lane line confidence of the current lane line.

[0006] In the above scheme, a two-dimensional lane line confidence score is calculated between each current lane line in the current lane line set and the historical lane line set, based on lane line distance confidence and lane line angle confidence in direction. Then, the current lane line set is matched with the historical lane line set based on the lane line confidence score set, transforming the single-frame lane lines in a discrete single frame into the tracking between historical and current lane lines in continuous frames, thus obtaining lane line tracking results. By simultaneously considering lane line distance confidence and lane line angle confidence, it avoids erroneous associations that may occur when matching based on a single feature, significantly improving the matching accuracy of lane lines between continuous frames. Finally, a stable matching and tracking of the current lane line set and the historical lane line set is completed based on the lane line confidence score set, resulting in a coherent and positionally smooth lane line tracking result, effectively eliminating inter-frame jumps and breaks, and enhancing the accuracy of lane line tracking in the autonomous driving system.

[0007] Further, obtaining the current lane line set of the road segment to be tracked and the historical lane line set corresponding to each current lane line in the current lane line set includes: Obtain the set of historical initial lane lines corresponding to each current lane line in the current lane line set; Perform a lane line screening operation on each of the historical initial lane line sets to obtain a lane line screening result set for each of the current lane lines. Then, based on the lane line screening result set, select the historical lane line set corresponding to each current lane line from the historical initial lane line set corresponding to each current lane line. The initial screening of lane lines includes: For any current lane line, based on the current lane line and each historical initial lane line in the corresponding historical initial lane line set, obtain a set of matching initial sampling point pairs between each historical initial lane line and the current lane line, determine a set of matching initial sampling point pair distances based on the set of matching initial sampling point pairs, and then obtain the initial lane line screening result between each historical initial lane line and the current lane line based on the second preset lane line distance threshold and the set of matching initial sampling point pair distances. Based on the current lane line and the corresponding set of historical initial lane lines, a set of lane line screening results is obtained for each current lane line.

[0008] In the above scheme, a coarse screening layer is added before fine matching by introducing a lane line initial screening process. It compares the distance set of the initial sampling points with a second preset lane line distance threshold to quickly eliminate obviously irrelevant historical initial lane lines from the historical initial lane line set, thereby significantly reducing the size of the historical lane line set for subsequent confidence calculation, thus improving lane line tracking efficiency and accuracy.

[0009] Further, for any current lane line, based on the current lane line and each historical lane line in its corresponding historical lane line set, obtaining the distance set of matching sampling point pairs and the angle set of matching sampling point pairs for each historical lane line includes: The current lane line sample pair set is obtained based on a preset sampling order; For any historical lane line in the set of historical lane lines, the set of historical lane line sampling points of the historical lane line is obtained based on a preset sampling order. Perform a sampling point matching operation on the historical lane line sampling point set to obtain a set of matching sampling point pairs between the historical lane line and the current lane line; The set of distances between the historical lane line and the current lane line and the set of included angles between the matching sampling point pairs are obtained from the set of matching sampling point pairs.

[0010] In the above scheme, the preset sampling order limits the current lane line sampling pair set and the historical lane line sampling point set to be along the lane line direction, thereby ensuring the smooth execution of the sampling point matching action, and thus obtaining the matching sampling point pair set of the historical lane line and the current lane line, and thus obtaining the accurate matching sampling point pair distance set and the matching sampling point pair angle set.

[0011] Further, the sampling point matching action is performed on the historical lane line sampling point set to obtain a set of matching sampling point pairs between the historical lane line and the current lane line. The sampling point matching action includes: For any historical lane line sampling point in the set of historical lane line sampling points, a corresponding current lane line sampling point is matched from the set of current lane line sampling points of the current lane line according to a preset extension direction, and then a matching sampling point pair is constructed based on the historical lane line sampling point and the corresponding current lane line sampling point. The set of matching sampling point pairs between the historical lane line and the current lane line is obtained based on all the matching sampling point pairs.

[0012] In the above scheme, the corresponding real-time lane line sampling point is found for each historical lane line sampling point on the historical lane line according to the preset extension direction. This ensures that the matching process is carried out in an orderly manner along the direction of the lane line, so that the final matching sampling point distance set can more accurately reflect the overall deviation of the two lane lines in space.

[0013] Further, the lane line distance confidence score of each historical lane line is obtained based on the distance set of the matching sampling points corresponding to each historical lane line and a first preset lane line distance threshold, including: Based on the set of distances between the matched sampling points, the distance between the old and new lane lines is obtained; The lane distance confidence level is obtained based on the first preset lane distance threshold, the distance between the new and old lane lines, and the preset lane distance confidence algorithm.

[0014] In the above scheme, the distance between the old and new lane lines is obtained by processing the distance set of the matching sampling points, and then substituted into the preset lane line distance confidence algorithm along with the first preset lane line distance threshold to obtain a lane line distance confidence score that can more accurately determine the degree of matching.

[0015] Further, the step of obtaining the lane line angle confidence level of each historical lane line based on the set of angles between the matching sampling points corresponding to each historical lane line and a preset lane line angle threshold includes: Select the angle with the largest value from the set of angles between the matching sampling points as the target angle between the matching sampling points; The lane line angle confidence is obtained based on the preset lane line angle threshold, the angle between the target matching sampling points, and the preset lane line angle confidence algorithm.

[0016] In the above scheme, the worst-case directional deviation is characterized by selecting the angle between the target matching sampling points (i.e., the maximum angle), and then substituting it together with the preset lane line angle threshold into the preset lane line angle confidence algorithm. This effectively filters out erroneous matching candidates caused by lane line bifurcation or curvature changes, thereby improving the accuracy of matching.

[0017] Further, the step of obtaining the lane line confidence level corresponding to each historical lane line based on the lane line distance confidence level and the lane line angle confidence level corresponding to each historical lane line, so as to obtain a subset of lane line confidence levels of the current lane line according to all the lane line confidence levels of the current lane line, includes: Based on the preset distance lower limit weight value, the preset distance upper limit weight value, and the lane line distance confidence value corresponding to each historical lane line, calculate the lane line distance confidence value corresponding to each historical lane line; The confidence weight value of the lane line angle corresponding to each historical lane line is determined based on the lane line distance confidence weight value; For each historical lane line, the lane line distance confidence weight value, lane line angle confidence weight value, lane line distance confidence value, and lane line angle confidence value are weighted and summed to obtain the lane line confidence value for each historical lane line. Based on all the lane line confidence values ​​of the current lane line, a subset of the lane line confidence values ​​for the current lane line is obtained.

[0018] In the above scheme, the lane line distance confidence weight value is dynamically calculated between the preset lower limit weight value and the preset upper limit weight value based on the lane line distance confidence value, and the lane line angle confidence weight value is determined accordingly. This achieves weighted summation, thereby effectively and adaptively adjusting the contribution ratio of distance and angle factors in the final lane line confidence value, making the matching strategy more flexible in different scenarios.

[0019] Another embodiment of the present invention provides a lane line tracking device, comprising: The basic data acquisition module is used to acquire the current lane line set of the road segment to be tracked and the historical lane line set corresponding to each current lane line in the current lane line set; The lane line confidence set acquisition module is used to perform a lane line confidence subset acquisition action for each of the historical lane line sets to obtain the lane line confidence subset of each current lane line, and then obtain the lane line confidence set corresponding to the current lane line set based on all the lane line confidence subsets; The lane line tracking result acquisition module is used to match the current lane line set with the historical lane line set based on the lane line confidence set to obtain the lane line tracking result of the road segment to be tracked. The lane line confidence subset acquisition action includes: For any current lane line, based on the current lane line and each historical lane line in the corresponding historical lane line set, obtain the set of distances between matching sampling points and the set of included angles between matching sampling points for each historical lane line; The lane line distance confidence level of each historical lane line is obtained by using the distance set of the matching sampling points corresponding to each historical lane line and the first preset lane line distance threshold. The confidence level of the lane line angle of each historical lane line is obtained based on the set of angles of the matching sampling points corresponding to each historical lane line and the preset lane line angle threshold. Based on the lane distance confidence and lane angle confidence of each historical lane line, the lane line confidence of each historical lane line is obtained, so as to obtain a subset of the lane line confidence of the current lane line according to all the lane line confidence of the current lane line.

[0020] Another embodiment of the present invention provides a terminal device, including: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements the steps of a lane line tracking method as described in the present invention.

[0021] Another embodiment of the present invention provides a computer-readable storage medium item, including: a stored computer program, which, when the computer program is running, controls the device where the computer-readable storage medium is located to perform steps of a lane line tracking method of the present invention. Attached Figure Description

[0022] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments 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 from these drawings without creative effort.

[0023] Figure 1 This is a schematic flowchart of a lane line tracking method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of a lane tracking device according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the distance set of matching sampling point pairs provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the set of included angles of matching sampling points provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of lane distance confidence provided in an embodiment of the present invention; Figure 6 This is a schematic diagram of the confidence level of the lane line angle provided in an embodiment of the present invention; Attached label: 1. Current lane line; 2. Historical lane line. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. 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.

[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0026] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0027] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0028] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0029] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0030] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.

[0031] See Figure 1 In order to solve the above-mentioned problems in the prior art, an embodiment of the present invention provides a lane line tracking method, device, terminal equipment and storage medium, which can solve the above problems and improve the accuracy of lane line tracking.

[0032] This invention provides a lane line tracking method, comprising: Step S1: Obtain the current lane line set of the road segment to be tracked and the historical lane line set corresponding to each current lane line in the current lane line set; Step S2: Perform a lane line confidence subset acquisition action for each of the historical lane line sets to obtain a lane line confidence subset for each current lane line, and then obtain the lane line confidence set corresponding to the current lane line set based on all the lane line confidence subsets; Step S3: Match the current lane line set with the historical lane line set based on the lane line confidence set to obtain the lane line tracking result of the road segment to be tracked; The lane line confidence subset acquisition action includes: For any current lane line, based on the current lane line and each historical lane line in the corresponding historical lane line set, obtain the set of distances between matching sampling points and the set of included angles between matching sampling points for each historical lane line; The lane line distance confidence level of each historical lane line is obtained by using the distance set of the matching sampling points corresponding to each historical lane line and the first preset lane line distance threshold. The confidence level of the lane line angle of each historical lane line is obtained based on the set of angles of the matching sampling points corresponding to each historical lane line and the preset lane line angle threshold. Based on the lane distance confidence and lane angle confidence of each historical lane line, the lane line confidence of each historical lane line is obtained, so as to obtain a subset of the lane line confidence of the current lane line according to all the lane line confidence of the current lane line.

[0033] In the above scheme, a two-dimensional lane line confidence score is calculated between each current lane line in the current lane line set and the historical lane line set, based on lane line distance confidence and lane line angle confidence in direction. Then, the current lane line set is matched with the historical lane line set based on the lane line confidence score set, transforming the single-frame lane lines in a discrete single frame into the tracking between historical and current lane lines in continuous frames, thus obtaining lane line tracking results. By simultaneously considering lane line distance confidence and lane line angle confidence, it avoids erroneous associations that may occur when matching based on a single feature, significantly improving the matching accuracy of lane lines between continuous frames. Finally, a stable matching and tracking of the current lane line set and the historical lane line set is completed based on the lane line confidence score set, resulting in a coherent and positionally smooth lane line tracking result, effectively eliminating inter-frame jumps and breaks, and enhancing the accuracy of lane line tracking in the autonomous driving system.

[0034] It should be noted that the current lane line set originates from the lane line structure information obtained by a perception model such as MapTR through vectorized perception of the current frame image, including the current lane line sampling point set for each current lane line; the historical lane line set is the lane line information that has been previously tracked and stored, which includes the historical lane line sampling point set, fitting parameters, and a unique tracking identifier for each historical lane line. Furthermore, for the received first frame of single-frame perception data, it is determined whether the lane line length is greater than a preset lane line length threshold (e.g., 3m) and whether the number of sampling points is greater than a preset sampling point number threshold (e.g., 3). When both features reach the threshold, it is determined to be a valid lane line, which needs to be tracked. A unique tracking number is assigned to the valid lane line, and the information of the first frame (point cloud, pose, coefficients) is stored in the historical tracking information to construct the initial tracking target. Subsequently, using the pose information, the current lane line is transformed to the historical frame coordinate system. The matching sampling point pair distance set and the matching sampling point pair angle set are calculated by fitting the curve with the historical lane line. It can be understood that the current perception result is already based on the vehicle coordinate system. At the same time, the vehicle will output its own positioning pose in the map (including orientation and position). Based on the positioning pose, the perception result and the historical lane line point cloud can be transformed to their respective corresponding vehicle coordinate systems.

[0035] In another embodiment, obtaining the current lane line set of the road segment to be tracked and the historical lane line set corresponding to each current lane line in the current lane line set includes: Obtain the set of historical initial lane lines corresponding to each current lane line in the current lane line set; Perform a lane line screening operation on each of the historical initial lane line sets to obtain a lane line screening result set for each of the current lane lines. Then, based on the lane line screening result set, select the historical lane line set corresponding to each current lane line from the historical initial lane line set corresponding to each current lane line. The initial screening of lane lines includes: For any current lane line, based on the current lane line and each historical initial lane line in the corresponding historical initial lane line set, obtain a set of matching initial sampling point pairs between each historical initial lane line and the current lane line, determine a set of matching initial sampling point pair distances based on the set of matching initial sampling point pairs, and then obtain the initial lane line screening result between each historical initial lane line and the current lane line based on the second preset lane line distance threshold and the set of matching initial sampling point pair distances. Based on the current lane line and the corresponding set of historical initial lane lines, a set of lane line screening results is obtained for each current lane line.

[0036] It should be noted that for each newly received frame of perception results, each lane line in the current frame is taken as the current lane line set, and it is matched with the historical lane line set in the historical tracking information. Understandably, for each current lane line in the current lane line set, the initial candidate for matching is the historical initial lane line set. Therefore, it is necessary to obtain the historical initial lane line set corresponding to each current lane line. To improve the accuracy of lane line tracking, spatial proximity can be used to perform a coarse matching of the historical initial lane line set to achieve preliminary filtering and obtain a candidate historical lane line set. This reduces the computational load when matching the historical lane line set with each current lane line in the subsequent process. Each frame of perception results may have multiple lane lines; therefore, the historical initial lane line set corresponding to each current lane line in the current lane line set is first obtained. The historical initial lane line set is the unfiltered set of lane lines. A lane line initial screening operation is performed on each of the historical initial lane line sets to obtain a lane line initial screening result set for each of the current lane lines. Then, based on the lane line initial screening result set, the historical lane line set corresponding to each current lane line is selected from the historical initial lane line set corresponding to each current lane line. The lane line initial screening operation includes: for any current lane line, obtaining a set of matching initial sampling point pairs between each historical initial lane line and the current lane line based on the current lane line and each historical initial lane line in its corresponding historical initial lane line set; determining a set of matching initial sampling point pair distances based on the set of matching initial sampling point pairs; and obtaining the lane line initial screening result for each historical initial lane line and the current lane line based on a second preset lane line distance threshold and the set of matching initial sampling point pair distances; and obtaining a lane line initial screening result set for each current lane line based on all the lane line initial screening results obtained from the current lane line and the corresponding historical initial lane line set.

[0037] What is understandable is that there is a certain functional relationship between the curve fitted by the current lane line sampling points and the set of distances for the initial matching sampling points. This functional relationship can be used to obtain the distance of each current lane line sampling point to the initial matching sampling point pair, thereby improving computational efficiency. Then, the distances of the initial matching sampling point pairs in the set of distances are averaged to obtain the average processed distance between the old and new lane lines. This can be used as the lateral distance between the geometric center of the historical initial lane line and the current lane line. What is understandable is that... Figure 3Taking the current lane line and historical lane line as examples to construct the coordinate axis, the direction perpendicular to historical lane line 2 is the Y-axis direction, the X-axis is the forward direction of the current lane line, and the Z-axis is the direction perpendicular to the XY plane. The lateral distance is the distance along the Y-axis. Based on the second preset lane line distance threshold and the average processing distance of the new and old lane lines, the initial lane line screening result for the historical initial lane line is obtained. If the average processing distance of the new and old lane lines is less than or equal to the second preset lane line distance threshold, the initial lane line screening result is considered acceptable; if the average processing distance of the new and old lane lines is greater than the second preset lane line distance threshold, the initial lane line screening result is considered unacceptable. Once the initial lane line screening result set for each current lane line in the current lane line set is obtained, the historical initial lane lines with unacceptable initial screening results are removed from each current lane line, forming the historical lane line set corresponding to each current lane line for subsequent more refined matching.

[0038] In another embodiment, for any current lane line, obtaining the set of distances between matching sample points and the set of included angles between matching sample points for each historical lane line in the set of historical lane lines and the current lane line itself, includes: The current lane line sample pair set is obtained based on a preset sampling order; For any historical lane line in the set of historical lane lines, the set of historical lane line sampling points of the historical lane line is obtained based on a preset sampling order. Perform a sampling point matching operation on the historical lane line sampling point set to obtain a set of matching sampling point pairs between the historical lane line and the current lane line; The set of distances between the historical lane line and the current lane line and the set of included angles between the matching sampling point pairs are obtained from the set of matching sampling point pairs.

[0039] It should be noted that the current lane line sampling pair set is obtained based on a preset sampling order. For any historical lane line in the historical lane line set, the historical lane line sampling point set is obtained based on a preset sampling order, which is the order based on the driving direction of the autonomous vehicle. A sampling point matching operation is performed on the historical lane line sampling point set to obtain the matching sampling point pair set between the historical lane line and the current lane line. The matching sampling point pair distance set is obtained by calculating the lateral distance between each matching sampling point pair. The matching sampling point pair angle set is obtained by calculating the angle between the local tangent direction of the current lane line and the historical lane line at each matching sampling point pair.

[0040] In another embodiment, the sampling point matching action performed on the historical lane line sampling point set to obtain a set of matching sampling point pairs between the historical lane line and the current lane line includes: For any historical lane line sampling point in the set of historical lane line sampling points, a corresponding current lane line sampling point is matched from the set of current lane line sampling points of the current lane line according to a preset extension direction, and then a matching sampling point pair is constructed based on the historical lane line sampling point and the corresponding current lane line sampling point. The set of matching sampling point pairs between the historical lane line and the current lane line is obtained based on all the matching sampling point pairs.

[0041] It should be noted that for any historical lane line sampling point in the historical lane line sampling point set, it is extended from a preset extension direction to obtain its intersection with the current lane line. The current lane line sampling point closest to the intersection point is then obtained as the current lane line sampling point corresponding to that historical lane line sampling point, and the two are thus considered as a pair of matched sampling points. It is understandable that because there is a time misalignment between the current lane line and the historical lane line set, if historical lane line sampling points are selected sequentially along the preset sampling order and the corresponding current lane line sampling points are obtained, the initial historical lane line sampling point will not match the corresponding current lane line sampling point until the first sampling point in the current lane line sampling point is matched along the lane line direction. Finally, a set of matched sampling point pairs between the historical lane line and the current lane line is obtained based on all the matched sampling point pairs. Then, the distance between the two sampling points of each matched sampling point pair is obtained based on the set of matched sampling point pairs, which is used as the set of matched sampling point pair distances. It is understandable that there is a certain functional relationship between the curve fitted by the current lane line sampling point and the set of distances between matching sampling point pairs. This functional relationship can be used to obtain the distance between matching sampling point pairs for each current lane line sampling point, thereby improving computational efficiency. Specifically, for each initial matching sampling point pair (or each current lane line sampling point), the formula for obtaining the distance between its corresponding matching sampling point pairs is as follows: ; in, This is the distance between the matching sample point pairs corresponding to the i-th initial matching sample point pair formed by the historical lane line and the current lane line (or the i-th current lane line sample point in the current lane line), i.e., as shown below. Figure 3 shown . This refers to the coordinate data (i.e., x-axis coordinates, where the x-axis represents the current lane's forward direction) of the i-th sampling point in the current lane line. c0, c1, and c2 are preset fitting coefficients, which can be calculated through a third-order fitting of historical lane line point clouds. It is understandable that this is based on... Figure 3 The red dashed line represents historical lane line 2, the black dashed line represents current lane line 1, and the green line represents the distance between matching sampling point pairs. Starting from the first current lane line sampling point of current lane line 1 and ending from the last historical lane line sampling point of historical lane line 2, if there are i pairs of matching sampling point pairs, then there are i pairs of matching sampling point pair distances.

[0042] Furthermore, the specific calculation method for the included angle of the matching sampling points is as follows: ; in, The angle between the (i-1)th matching sampling point and the other side. This refers to the coordinate data (i.e., x-axis coordinate data) of the i-th sampling point of the current lane line. This refers to the coordinate data (i.e., x-axis coordinate data) of the (i-1)th sampling point of the current lane line. The distance between the i-th matching sampling point pair The distance between the (i-1)th matching sampling point pair. Specifically, the coordinate difference data between any two adjacent historical lane line sampling points in the historical lane line sampling point set is obtained to acquire the historical sampling point coordinate difference data set, which is the method for obtaining... and The difference; based on the distance set of the matched sampling point pairs and the coordinate difference data set of the historical sampling points, the angle set of the matched sampling point pairs is obtained, which is based on This yields the set of included angles between the matched sampling points. (Reference) Figure 4 What I can understand is... Figure 4 and Figure 3 Similarly, the red dashed line represents the historical lane line, the black dashed line represents the current lane line, and the green line represents the distance between matching sampling point pairs. Assuming there are i current lane line sampling points, there are i-1 matching sampling point pairs with included angles.

[0043] In another embodiment, obtaining the lane line distance confidence score of each historical lane line based on the distance set of the matching sampling points corresponding to each historical lane line and a first preset lane line distance threshold includes: Based on the set of distances between the matched sampling points, the distance between the old and new lane lines is obtained; The lane distance confidence level is obtained based on the first preset lane distance threshold, the distance between the new and old lane lines, and the preset lane distance confidence algorithm.

[0044] It should be noted that the preset lane line distance confidence algorithm is specifically as follows: ; Where confidenceD is the lane distance confidence score, avgD is the distance between the old and new lane lines (obtained by averaging the distances of the matched sampling points), and distanceTh is the first preset lane distance threshold. The corresponding lane distance confidence curve is shown below. Figure 5 As shown, as the average distance between the old and new lane lines increases, the lane line distance confidence curve decreases smoothly and rapidly. Preferably, through debugging using perceived lane line data recorded by real vehicles, the first preset lane line distance threshold distanceTh is preferably 0.6m, and can be further adjusted and optimized as the amount of data increases.

[0045] In another embodiment, obtaining the lane line angle confidence level of each historical lane line based on the set of angles between the matching sampling points corresponding to each historical lane line and a preset lane line angle threshold includes: Select the angle with the largest value from the set of angles between the matching sampling points as the target angle between the matching sampling points; The lane line angle confidence is obtained based on the preset lane line angle threshold, the angle between the target matching sampling points, and the preset lane line angle confidence algorithm.

[0046] It should be noted that the preset lane line angle confidence algorithm is as follows: ; Where confidenceA is the lane angle confidence score, maxA is the angle between the target matching sample points (i.e., the angle between the matching sample points with the largest value in the set of matching sample point angles), and angleTh is a preset lane angle threshold. When maxA is 0, the lane angle confidence score is 0. The corresponding lane angle confidence score curve is shown below. Figure 6 As shown, it is understandable that in practical applications, maxA=0 indicates that the number of points from which the input lane line participates in the calculation of the matching sampling points for the included angle is insufficient. Therefore, the confidence level of the lane line included angle should be 0 in this case. Figure 6 The lane line angle confidence curve shows that, during the effective calculation of the angle, as the angle between the target matching sampling point and the target gradually increases, the lane line angle confidence smoothly and rapidly decreases near the preset lane line angle threshold. Preferably, through debugging using perceived lane line data recorded by real vehicles, the preset lane line angle threshold angleTh is preferably 2.5 degrees, and can be further adjusted and optimized as the amount of data increases.

[0047] In another embodiment, the step of obtaining the lane line confidence level corresponding to each historical lane line based on the lane line distance confidence level and the lane line angle confidence level corresponding to each historical lane line, and obtaining a subset of the lane line confidence levels of the current lane line based on all the lane line confidence levels of the current lane line, includes: Based on the preset distance lower limit weight value, the preset distance upper limit weight value, and the lane line distance confidence value corresponding to each historical lane line, calculate the lane line distance confidence value corresponding to each historical lane line; The confidence weight value of the lane line angle corresponding to each historical lane line is determined based on the lane line distance confidence weight value; For each historical lane line, the lane line distance confidence weight value, lane line angle confidence weight value, lane line distance confidence value, and lane line angle confidence value are weighted and summed to obtain the lane line confidence value for each historical lane line. Based on all the lane line confidence values ​​of the current lane line, a subset of the lane line confidence values ​​for the current lane line is obtained.

[0048] It should be noted that when both the lane distance confidence score and the lane angle confidence score are small, it indicates that the current lane line's distance and angle with the historical lane line are both large, and the final lane line confidence score should be small. When the lane distance confidence score is large, but the lane angle confidence score is small, it indicates that the current lane line's lateral distance with the historical lane line is small, but the angle is large, suggesting a tendency for the lane lines to fork. In this case, the lane distance confidence score weight should be decreased, and the lane angle confidence score weight should be increased, allowing the lane angle confidence score weight to dominate the final confidence score. When the lane distance confidence score is small, but the lane angle confidence score is large, it indicates that the current lane line's lateral distance with the historical lane line is large, and the angle is small, suggesting that the current lane line may be a parallel, non-matching lane line. In this case, the lane angle confidence score weight should be decreased, and the lane distance confidence score weight should be increased. In summary, the final lane line confidence score calculation formula is: ; ; Where confidence is the lane line confidence level, weightD is the lane line distance confidence level weight, and weightA is the lane line angle confidence level weight. K is the preset upper limit weight value, and H is the preset lower limit weight value. Preferably, K is 0.2 and H is 0.3, so that the lane line distance confidence level weight can be dynamically adjusted between 0.3 and 0.5.

[0049] Furthermore, based on lane line confidence and a confidence threshold, the lane line tracking results for the current frame and historical frames of the tracked road segment are obtained: lane lines with confidence scores higher than the confidence threshold are matched, i.e., the current lane line is associated with a matched historical lane line; otherwise, a new historical lane line is generated. For successfully matched lane line tracking, the current frame point cloud and the historical frame point cloud are merged, polynomial refitting is performed, and adaptive sampling is used to generate smooth and sparse trajectory points, achieving temporal smoothing and enhancement, and updating the historical lane line information and its topological relationships. This process uses existing lane line point cloud fitting methods, such as cubic polynomial curve fitting. After obtaining the lane line fitting parameters, the lane lines can be resampled according to the parameters, and then the updated historical tracking information is published to the current frame pose coordinates. After obtaining the lane line tracking results, a map is constructed based on the lane line tracking results and then used for subsequent map-based autonomous driving.

[0050] like Figure 2 As shown, based on the above method embodiments, corresponding apparatus embodiments are provided; An embodiment of the present invention provides a lane line tracking device, comprising: The basic data acquisition module is used to acquire the current lane line set of the road segment to be tracked and the historical lane line set corresponding to each current lane line in the current lane line set; The lane line confidence set acquisition module is used to perform a lane line confidence subset acquisition action for each of the historical lane line sets to obtain the lane line confidence subset of each current lane line, and then obtain the lane line confidence set corresponding to the current lane line set based on all the lane line confidence subsets; The lane line tracking result acquisition module is used to match the current lane line set with the historical lane line set based on the lane line confidence set to obtain the lane line tracking result of the road segment to be tracked. The lane line confidence subset acquisition action includes: For any current lane line, based on the current lane line and each historical lane line in the corresponding historical lane line set, obtain the set of distances between matching sampling points and the set of included angles between matching sampling points for each historical lane line; The lane line distance confidence level of each historical lane line is obtained by using the distance set of the matching sampling points corresponding to each historical lane line and the first preset lane line distance threshold. The confidence level of the lane line angle of each historical lane line is obtained based on the set of angles of the matching sampling points corresponding to each historical lane line and the preset lane line angle threshold. Based on the lane distance confidence and lane angle confidence of each historical lane line, the lane line confidence of each historical lane line is obtained, so as to obtain a subset of the lane line confidence of the current lane line according to all the lane line confidence of the current lane line.

[0051] It is understood that the above-described device embodiments correspond to the method embodiments of the present invention, and can implement the lane line tracking method provided by any of the above-described method embodiments of the present invention.

[0052] It should be noted that the device embodiments described above are merely illustrative, and some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can specifically be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0053] Based on the above-described embodiment of the lane line tracking method, another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements a lane line tracking method according to any embodiment of the present invention.

[0054] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0055] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0056] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0057] Based on the above-described method embodiments, another embodiment of the present invention provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute a lane line tracking method as described in any of the above-described method embodiments of the present invention.

[0058] The modules / units integrated in the device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0059] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A lane line tracking method, characterized in that, include: Obtain the current lane line set of the road segment to be tracked and the historical lane line set corresponding to each current lane line in the current lane line set; For each of the historical lane line sets, a lane line confidence subset acquisition action is performed to obtain a lane line confidence subset for each of the current lane lines. Then, based on all the lane line confidence subsets, a lane line confidence set corresponding to the current lane line set is obtained. Based on the lane line confidence set, the current lane line set is matched with the historical lane line set to obtain the lane line tracking result of the road segment to be tracked; The lane line confidence subset acquisition action includes: For any current lane line, based on the current lane line and each historical lane line in the corresponding historical lane line set, obtain the set of distances between matching sampling points and the set of included angles between matching sampling points for each historical lane line; The lane line distance confidence level of each historical lane line is obtained by the matching sampling point pair distance set corresponding to each historical lane line and the first preset lane line distance threshold. The confidence level of the lane line angle of each historical lane line is obtained based on the set of angles of the matching sampling points corresponding to each historical lane line and a preset lane line angle threshold. Based on the lane distance confidence and lane angle confidence of each historical lane line, the lane line confidence of each historical lane line is obtained, so as to obtain a subset of the lane line confidence of the current lane line according to all the lane line confidence of the current lane line.

2. The lane line tracking method according to claim 1, characterized in that, The step of obtaining the current lane line set of the road segment to be tracked and the historical lane line set corresponding to each current lane line in the current lane line set includes: Obtain the set of historical initial lane lines corresponding to each current lane line in the current lane line set; Perform a lane line screening operation on each of the historical initial lane line sets to obtain a lane line screening result set for each of the current lane lines. Then, based on the lane line screening result set, select the historical lane line set corresponding to each current lane line from the historical initial lane line set corresponding to each current lane line. The initial screening of lane lines includes: For any current lane line, based on the current lane line and each historical initial lane line in the corresponding historical initial lane line set, obtain a set of matching initial sampling point pairs between each historical initial lane line and the current lane line, determine a set of matching initial sampling point pair distances based on the set of matching initial sampling point pairs, and then obtain the initial lane line screening result between each historical initial lane line and the current lane line based on the second preset lane line distance threshold and the set of matching initial sampling point pair distances. Based on the current lane line and the corresponding set of historical initial lane lines, a set of lane line screening results is obtained for each current lane line.

3. The lane line tracking method according to claim 2, characterized in that, For any given current lane line, based on the current lane line and each historical lane line in its corresponding historical lane line set, the set of distances between matching sample points and the set of included angles between matching sample points for each historical lane line are obtained, including: The current lane line sample pair set is obtained based on a preset sampling order; For any historical lane line in the set of historical lane lines, the set of historical lane line sampling points of the historical lane line is obtained based on a preset sampling order. Perform a sampling point matching operation on the historical lane line sampling point set to obtain a set of matching sampling point pairs between the historical lane line and the current lane line; The set of distances between the historical lane line and the current lane line and the set of included angles between the matching sampling point pairs are obtained from the set of matching sampling point pairs.

4. The lane line tracking method according to claim 3, characterized in that, The step of performing a sampling point matching operation on the historical lane line sampling point set to obtain a set of matching sampling point pairs between the historical lane line and the current lane line includes: For any historical lane line sampling point in the set of historical lane line sampling points, a corresponding current lane line sampling point is matched from the set of current lane line sampling points of the current lane line according to a preset extension direction, and then a matching sampling point pair is constructed based on the historical lane line sampling point and the corresponding current lane line sampling point. The set of matching sampling point pairs between the historical lane line and the current lane line is obtained based on all the matching sampling point pairs.

5. The lane line tracking method according to claim 1, characterized in that, The step of obtaining the lane line distance confidence score of each historical lane line based on the distance set of the matching sampling points corresponding to each historical lane line and the first preset lane line distance threshold includes: Based on the set of distances between the matched sampling points, the distance between the old and new lane lines is obtained; The lane distance confidence level is obtained based on the first preset lane distance threshold, the distance between the new and old lane lines, and the preset lane distance confidence algorithm.

6. The lane line tracking method according to claim 1, characterized in that, The process of obtaining the lane angle confidence level of each historical lane line based on the set of angles between the matching sampling points corresponding to each historical lane line and a preset lane angle threshold includes: Select the angle with the largest value from the set of angles between the matching sampling points as the target angle between the matching sampling points; The lane line angle confidence is obtained based on the preset lane line angle threshold, the angle between the target matching sampling points, and the preset lane line angle confidence algorithm.

7. A lane line tracking method according to any one of claims 1-6, characterized in that, The process of obtaining the lane line confidence level for each historical lane line based on the lane line distance confidence level and the lane line angle confidence level, and then obtaining a subset of the lane line confidence levels for the current lane line based on all the lane line confidence levels of the current lane line, includes: Based on the preset distance lower limit weight value, the preset distance upper limit weight value, and the lane line distance confidence value corresponding to each historical lane line, calculate the lane line distance confidence value corresponding to each historical lane line; The confidence weight value of the lane line angle corresponding to each historical lane line is determined based on the lane line distance confidence weight value; For each historical lane line, the lane line distance confidence weight value, lane line angle confidence weight value, lane line distance confidence value, and lane line angle confidence value are weighted and summed to obtain the lane line confidence value for each historical lane line. Based on all the lane line confidence values ​​of the current lane line, a subset of the lane line confidence values ​​for the current lane line is obtained.

8. A lane line tracking device, characterized in that, include: The basic data acquisition module is used to acquire the current lane line set of the road segment to be tracked and the historical lane line set corresponding to each current lane line in the current lane line set; The lane line confidence set acquisition module is used to perform a lane line confidence subset acquisition action for each of the historical lane line sets to obtain the lane line confidence subset of each current lane line, and then obtain the lane line confidence set corresponding to the current lane line set based on all the lane line confidence subsets; The lane line tracking result acquisition module is used to match the current lane line set with the historical lane line set based on the lane line confidence set to obtain the lane line tracking result of the road segment to be tracked. The lane line confidence subset acquisition action includes: For any current lane line, based on the current lane line and each historical lane line in the corresponding historical lane line set, obtain the set of distances between matching sampling points and the set of included angles between matching sampling points for each historical lane line; The lane line distance confidence level of each historical lane line is obtained by the matching sampling point pair distance set corresponding to each historical lane line and the first preset lane line distance threshold. The confidence level of the lane line angle of each historical lane line is obtained based on the set of angles of the matching sampling points corresponding to each historical lane line and a preset lane line angle threshold. Based on the lane distance confidence and lane angle confidence of each historical lane line, the lane line confidence of each historical lane line is obtained, so as to obtain a subset of the lane line confidence of the current lane line according to all the lane line confidence of the current lane line.

9. A terminal device, characterized in that, The system includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements a lane tracking method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, include: A stored computer program, wherein, when the computer program is executed, it controls the device containing the computer-readable storage medium to perform a lane line tracking method as described in any one of claims 1-7.

Images