Lane center line detection method and device, lane line detection apparatus, and storage medium

Abstract

The specification discloses a lane center line detection method and device, a lane line detection apparatus and a storage medium. A plurality of first discrete points are determined on the lane center line, the radii corresponding to the plurality of first discrete points are further obtained, and the radii corresponding to the plurality of first discrete points are compared with a preset radius threshold, so that the positional relationship between the lane center line and the lane reference line is detected according to the comparison result of the radii corresponding to the plurality of first discrete points and the preset radius threshold, to detect whether the lane center line is smooth. Based on the lane center line detection method, the lane center line can be judged, the lane center line prone to false detection can be detected, and the accuracy of the lane center line detection result is improved.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, and in particular to a method, apparatus, lane line detection device, and storage medium for lane center line detection. Background Technology

[0002] In the field of trajectory planning for autonomous driving, since vehicles need to generate travel trajectories based on lane centerlines to achieve autonomous driving on public urban roads, trajectory planning methods require a certain degree of smoothness in the lane centerlines. In particular, because an uneven lane centerline can affect the autonomous driving status of the vehicle, it is necessary to detect whether the lane centerline is smooth.

[0003] In related technologies, crowdsourced manual screening is used to find uneven lane centerlines. However, the accuracy of lane centerline detection using crowdsourced manual screening needs improvement. Summary of the Invention

[0004] The embodiments described in this specification aim to at least partially solve one of the technical problems in the related art. To this end, the embodiments described in this specification propose a lane centerline detection method, apparatus, lane line detection device, and storage medium.

[0005] This specification provides a lane centerline detection method, the method comprising: determining a plurality of first discrete points on the lane centerline; obtaining the radius corresponding to the plurality of first discrete points; wherein the radius is the radius of a first circle jointly determined by the first discrete point and two adjacent discrete points adjacent to the first discrete point; detecting the positional relationship between the lane centerline and a lane reference line based on a comparison result between the radius corresponding to the plurality of first discrete points and a preset radius threshold, so as to detect whether the lane centerline is smooth; wherein the lane reference line is used to represent the lane in which the lane centerline is located.

[0006] This specification provides a lane centerline detection device, comprising: a discrete point determination module for determining a plurality of first discrete points on the lane centerline; a radius acquisition module for acquiring the radius corresponding to the plurality of first discrete points; wherein the radius is the radius of a first circle jointly determined by the first discrete point and two adjacent discrete points adjacent to the first discrete point; and a positional relationship detection module for detecting the positional relationship between the lane centerline and a lane reference line based on a comparison result between the radius corresponding to the plurality of first discrete points and a preset radius threshold, so as to detect whether the lane centerline is smooth; wherein the lane reference line is used to represent the lane in which the lane centerline is located.

[0007] This specification provides a lane line detection device, the lane line detection including: a memory, and one or more processors communicatively connected to the memory; the memory stores instructions executable by the one or more processors, the instructions being executed by the one or more processors to cause the one or more processors to perform the steps of the method described in any of the above embodiments.

[0008] This specification provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in any of the above embodiments.

[0009] This specification provides a computer program product that includes instructions that, when executed by a processor of a computer device, enable the computer device to perform the steps of the method described in any of the above embodiments.

[0010] In the above-described embodiment, multiple first discrete points are determined on the lane centerline, and the radii corresponding to these points are further obtained. These radii are then compared with a preset radius threshold. Based on this comparison, the positional relationship between the lane centerline and the lane reference line is detected. The detected positional relationship indicates whether the lane centerline and the lane reference line tend to be parallel, thus determining whether the lane centerline is smooth. This lane centerline detection method can accurately detect lane centerlines that are prone to false detection (such as curves and irregular lanes), improving the accuracy of lane centerline detection results. Attached Figure Description

[0011] Figure 1a This is a schematic diagram of the lane centerline at a turn, provided for an embodiment of this specification.

[0012] Figure 1b A schematic diagram of the lane centerline of an irregular lane provided for the implementation of this specification.

[0013] Figure 1c This diagram illustrates an application scenario of the lane centerline detection method provided in the embodiments of this specification.

[0014] Figure 1d This is a flowchart illustrating the lane centerline detection method provided in the embodiments of this specification.

[0015] Figure 1e A schematic diagram of the lane centerline provided for embodiments of this specification.

[0016] Figure 1f A schematic diagram of the lane centerline provided for embodiments of this specification.

[0017] Figure 1g A schematic diagram showing the distribution of the first discrete point on the lane centerline provided for the implementation of this specification.

[0018] Figure 2 This is a flowchart illustrating the lane centerline detection method provided in the embodiments of this specification.

[0019] Figure 3 This is a flowchart illustrating the lane centerline detection method provided in the embodiments of this specification.

[0020] Figure 4 This is a flowchart illustrating the lane centerline detection method provided in the embodiments of this specification.

[0021] Figure 5 This is a flowchart illustrating the lane centerline detection method provided in the embodiments of this specification.

[0022] Figure 6 This is a flowchart illustrating the lane centerline detection method provided in the embodiments of this specification.

[0023] Figure 7 This is a flowchart illustrating the lane centerline detection method provided in the embodiments of this specification.

[0024] Figure 8 This is a flowchart illustrating the lane centerline detection method provided in the embodiments of this specification.

[0025] Figure 9 This is a schematic diagram of the lane centerline detection device provided in the embodiments of this specification. Detailed Implementation

[0026] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0027] In related technologies, uneven lane centerlines can be found through crowdsourced manual screening. Lane centerlines can also be detected using angle filtering or lane line distance methods.

[0028] However, crowdsourced manual screening methods in related technologies may result in missed detections. Please refer to... Figure 1a Angle-based filtering methods may falsely detect lane centerlines at curves. Please refer to... Figure 1b Methods based on lane line distance are easily affected by irregular lanes.

[0029] Therefore, this specification provides a lane centerline detection method by determining multiple first discrete points on the lane centerline and obtaining the radii corresponding to these first discrete points. The radius is the radius of a first circle determined by the first discrete point and two adjacent discrete points. Based on a comparison of the radii corresponding to the multiple first discrete points with a preset radius threshold, the positional relationship between the lane centerline and the lane reference line is detected. By determining whether the lane centerline and the lane reference line tend to be parallel based on the detected positional relationship, the smoothness of the lane centerline can be determined. Unsmooth lane centerlines can be filtered out, reducing the impact of lane curves and irregular lane lines on the smoothness detection of the lane centerline and improving the accuracy of lane centerline detection.

[0030] Please see Figure 1c The lane centerline detection method provided in this specification can be applied to... Figure 1c In vehicle 110, or applied to other devices that have the function of controlling vehicles (such as...) Figure 1c The system includes cloud servers (120), mobile terminals (130, etc.). The vehicle can be an autonomous vehicle, which can have partial or full autonomous driving capabilities. Specifically, the level of autonomous driving can be classified according to the Society of Automotive Engineers (SAE) standards: no automation (L0), driver assistance (L1), partial automation (L2), conditional automation (L3), high automation (L4), or full automation (L5). Vehicles or other equipment can implement this lane centerline detection method through their components (including hardware and software).

[0031] It is understood that the vehicle can be any of the following: car, truck, motorcycle, bus, recreational vehicle, amusement park vehicle, construction equipment (such as engineering vehicle), tram, golf cart, train, and handcart, etc., and the embodiments described in this specification do not particularly limit this.

[0032] It should be noted that after detecting the lane centerline using the lane centerline detection method provided in this specification, on the one hand, uneven lane centerlines can be repaired (e.g., automatically or manually) to meet the requirements of autonomous vehicle trajectory planning; on the other hand, uneven lane centerlines can be directly discarded, and new lane centerlines can be generated based on actual lane data to obtain smooth lane centerlines. Finally, the smooth lane centerlines are used to generate the autonomous vehicle's trajectory, ensuring the normal operation of the autonomous vehicle. For example, smooth lane centerlines can be applied to navigation maps (e.g., high-precision maps) to plan the vehicle's trajectory.

[0033] This specification provides a lane centerline detection method. Please refer to [link / reference]. Figure 1d The lane centerline detection method may include the following steps:

[0034] S110, Determine multiple first discrete points on the center line of the lane.

[0035] Please refer to Figure 1e In the diagram, A and B are the lane dividing lines of the lane where the vehicle is located, and C is the lane centerline, which can be understood as a virtual centerline at the center of the lane where the vehicle is located. For example, the distance between lane dividing lines A and B is denoted as 'd', and the distances between lane centerline C and lane dividing lines A and D are 'd / 2' respectively. In this embodiment, the lane centerline can be a line segment or an arc (or a curved segment). Please refer to [link / reference]. Figure 1f The lane centerline can also be generated by connecting multiple adjacent arcs or line segments.

[0036] The multiple first discrete points can be a series of discontinuous points located on the lane centerline. These multiple first discrete points can be evenly distributed along the lane centerline at equal intervals, or they can be distributed at unequal intervals.

[0037] Specifically, in one embodiment, the lane centerline can be discretized to obtain multiple first discrete points. In another embodiment, if a number of first discrete points on the lane centerline are pre-stored, then the multiple first discrete points on the lane centerline can be directly obtained.

[0038] S120. Obtain the radii corresponding to multiple first discrete points.

[0039] Wherein, the radius is the radius of the first circle determined by the first discrete point and the two adjacent discrete points adjacent to the first discrete point. For example, please refer to... Figure 1gThe first discrete points distributed on the lane centerline C include X, Y, M, N, P, and Q. Since X and Q are located at opposite ends of the lane centerline, the two adjacent discrete points of the first discrete point X can be Y and M. The two adjacent discrete points of the first discrete point Q can be P and N. The first discrete point Y, located in the middle of the lane centerline, can be adjacent to X and M. The first discrete point M, located in the middle of the lane centerline, can be adjacent to Y and N.

[0040] Specifically, when three points are not on the same straight line, a triangle is formed by these three points, and this triangle has one and only one circumcircle. That is, a circle can be determined based on three points not on the same straight line. When three points are on the same straight line, a line segment passing through these three points is determined, and a circle can be determined using this line segment as the straight line. Therefore, for any first discrete point on the lane centerline, this first discrete point and its two adjacent discrete points can jointly determine a circle, denoted as the first circle. The radius of the first circle is the radius corresponding to this arbitrary first discrete point.

[0041] S130. Based on the comparison results of the radii corresponding to multiple first discrete points and the preset radius threshold, detect the positional relationship between the lane centerline and the lane reference line to detect whether the lane centerline is smooth.

[0042] The lane reference line is used to indicate the lane in which the lane centerline is located. The lane reference line can be determined based on at least one of the lane dividing lines on both sides of the lane. The lane reference line can be understood as a virtual line representing the lane. The preset radius threshold can be a critical value used to determine whether the lane centerline is smooth. The preset radius threshold can be set based on the actual road conditions in the city. The preset radius threshold can be set based on the city level of the road. The preset radius threshold can be set based on the geographical features (e.g., hills, plains, basins) of the geographical area where the lane is located. The preset radius threshold can also be set based on the road level and administrative grade of a road. The positional relationship between the lane centerline and the lane reference line includes intersection and parallelism. In this embodiment, it is necessary to filter lane centerlines that intersect with the lane reference line and retain lane centerlines that tend to be parallel to the lane reference line. For example, the preset radius threshold can be any one of 1m, 0.8m, 1.2m, etc.

[0043] Specifically, multiple discrete points are distributed along the lane centerline, each with its own radius. A preset radius threshold is set in advance based on the actual situation. The radius corresponding to each discrete point is compared with the preset radius threshold. If the radius corresponding to the first discrete point is not less than the preset radius threshold, the position of the first discrete point can be determined to be smooth. If the radius corresponding to the first discrete point is less than the preset radius threshold, it indicates that the position of the first discrete point may not be smooth, requiring further judgment. Therefore, it is necessary to combine the comparison results of the radii corresponding to each first discrete point with the preset radius threshold to detect the positional relationship between the lane centerline and the lane reference line, thereby detecting whether the lane centerline is smooth. For example, based on the comparison results of the radii corresponding to multiple first discrete points with the preset radius threshold, if it is determined that the lane centerline and the lane reference line tend to be parallel, the lane centerline is determined to be smooth; if it is determined that the lane centerline intersects with the lane reference line, the lane centerline is determined to be unsmooth. In the aforementioned lane centerline detection method, multiple first discrete points are determined on the lane centerline, and the radii corresponding to these first discrete points are further obtained. These radii are then compared with a preset radius threshold. Based on this comparison, the positional relationship between the lane centerline and the lane reference line is determined. By analyzing the detected positional relationship, it can be determined whether the lane centerline and the lane reference line tend to be parallel, thus determining whether the lane centerline is smooth. This lane centerline detection method can be used to judge sequential lane centerlines, and can also detect lane centerlines that are prone to false positives or false negatives (such as curves and irregular lanes) as needed, reducing the probability of missed detections and improving the accuracy of lane centerline detection results.

[0044] In some implementations, please refer to Figure 2 Based on the comparison results between the radii corresponding to multiple first discrete points and a preset radius threshold, detecting the positional relationship between the lane centerline and the lane reference line may include the following steps:

[0045] S210. If there is a target radius smaller than a preset radius threshold among the radii corresponding to multiple first discrete points, at least two first target discrete points are determined among the multiple first discrete points on the lane centerline.

[0046] The first target discrete point can be any two of the first discrete points corresponding to the target radius, or it can be two first discrete points at a specified position on the lane centerline, or it can be the first discrete points at both ends of the lane centerline. Each first target discrete point corresponds to a target circle center. The target circle center can be the center of a circle determined based on the first target discrete point and two adjacent first discrete points.

[0047] Specifically, the radius corresponding to the first discrete point is compared with a preset radius threshold. If the radius corresponding to at least one first discrete point is smaller than the preset radius threshold, the radius corresponding to that at least one first discrete point is determined as the target radius. Furthermore, among the multiple first discrete points on the lane centerline, at least two first target discrete points are determined.

[0048] S220. Based on the positional relationship between the lane centerline and the target circle center, detect the positional relationship between the lane centerline and the lane reference line.

[0049] The positional relationship between the lane centerline and the target center includes being on the same side of the lane centerline or on different sides of the lane centerline.

[0050] Specifically, based on a first target discrete point and two adjacent first discrete points, a circle can be defined, and the center of this circle is denoted as the target circle center. The positional relationship between the target circle center and the lane centerline is determined, and this relationship is used to further determine the positional relationship between the lane centerline and the lane reference line. For example, if the target circle center is located on the same side of the lane centerline, it can be determined that the lane centerline and the lane reference line tend to be parallel, thus indicating a smooth lane centerline. If the target circle center is located on different sides of the lane centerline, it can be determined that the lane centerline and the lane reference line tend to intersect, thus indicating a non-smooth lane centerline.

[0051] In the above lane centerline detection method, at least two first target discrete points are determined by multiple first discrete points on the lane centerline, and the positional relationship between the lane centerline and the lane reference line is detected based on the positional relationship between the lane centerline and the target circle center corresponding to the first target discrete points. This reduces the impact of irregular lanes on lane centerline detection and improves detection accuracy.

[0052] In some implementations, please refer to Figure 3 Among multiple first discrete points on the lane centerline, at least two first target discrete points are determined, including: determining the first discrete points located at both ends of the lane centerline as the first target discrete points.

[0053] Specifically, if among the radii corresponding to multiple first discrete points there exists a target radius smaller than a preset radius threshold, the first discrete points located at both ends of the lane centerline (e.g., Figure 3 Points A and B at either end of the lane are designated as the first target discrete points. Based on these first target discrete points, and two adjacent first target discrete points on the lane centerline, a circle can be defined, with its center denoted as the target circle center. Based on the positional relationship between the target circle center and the lane centerline, it is determined whether the lane centerline and the lane reference line tend to be parallel or intersecting.

[0054] In the above lane centerline detection method, the difference between the trend of the lane centerline and the trend of the lane reference line is detected by the positional relationship between the target circle corresponding to the first discrete point at both ends of the lane centerline and the lane centerline, thereby reducing the detection accuracy of whether the lane centerline is smooth for irregular lanes.

[0055] In some implementations, the positional relationship between the lane centerline and the lane reference line is detected based on the positional relationship between the lane centerline and at least two target circle centers, in order to detect whether the lane centerline is smooth, including: if at least two target circle centers are located on the same side of the lane centerline, it is determined that the lane centerline and the lane reference line tend to be parallel.

[0056] Specifically, the positional relationship between the lane centerline and the lane reference line is further determined by the positional relationship between the target circle center and the lane centerline. When at least two target circle centers are located on the same side of the lane centerline, it is determined that the lane centerline and the lane reference line tend to be parallel, and thus the lane centerline can be considered smooth.

[0057] In the above-mentioned lane centerline detection method, the trend of the lane centerline is determined to be the same as the trend of the lane reference line by having at least two target circles of the vehicle located on the same side of the lane centerline, thereby improving the accuracy of lane centerline detection.

[0058] In some implementations, the lane to which the lane centerline belongs is provided with lane boundaries. The positional relationship between the lane centerline and a lane reference line is detected based on the positional relationship between the lane centerline and at least two target circles, including: if the at least two target circles are located on different sides of the lane centerline, detecting whether the lane centerline and the lane reference line tend to be parallel based on the positional relationship between the lane centerline and the lane boundaries.

[0059] Lane boundaries include lane dividing lines and / or lane boundary lines. Lane dividing lines are traffic markings used to separate traffic flows traveling in the same direction, typically white dashed / solid lines or yellow dashed / solid lines. Lane boundary lines are the edges of the lanes on either side, also known as lane edge lines. Specifically, the positional relationship between the lane centerline and the lane reference line is further determined by the positional relationship between the target circle's center and the lane centerline. When at least two target circles are located on different sides of the lane centerline, it is necessary to further determine the positional relationship between the lane centerline and the lane boundary lines to check whether the lane centerline and the lane reference line tend to be parallel. For example, if the lane centerline and lane boundary lines tend to be parallel, it can be determined that the lane centerline and lane reference line tend to be parallel, thus indicating that the lane centerline is smooth; if the lane centerline and lane boundary lines tend to intersect, it can be determined that the lane centerline and lane reference line tend to intersect, thus indicating that the lane centerline is not smooth.

[0060] In the above-mentioned lane centerline detection method, when at least two target circles are located on different sides of the lane centerline, the method further utilizes the positional relationship between the lane centerline and the lane boundary line to detect whether the trends between the lane centerline and the lane reference line are the same, thereby improving the accuracy of lane centerline detection.

[0061] In some implementations, the lane boundary includes a target line segment corresponding to the position of the lane centerline, with each end of the target line segment having a second discrete point. See also... Figure 4 Based on the positional relationship between the lane centerline and the lane boundary line, detecting whether the lane centerline and the lane reference line tend to be parallel can include the following steps:

[0062] S410. Based on the second discrete point, the second circle is jointly determined by two adjacent discrete points adjacent to the second discrete point.

[0063] In this embodiment, lane boundaries can correspond to multiple lane center lines connected end-to-end. Each lane center line has a portion of the lane boundary line at a corresponding position; this portion of the lane boundary line is called the target line segment. This target line segment can be discretized to obtain several corresponding discrete points. The discrete points located at both ends of the target line segment are denoted as the second discrete points. It should be noted that in this embodiment, the target line segment can be a line segment that is nearly parallel to a straight line, or it can be a curved segment.

[0064] In some cases, if at least two target circle centers are located on different sides of the lane centerline, it is necessary to further detect whether the lane centerline and lane reference line tend to be parallel based on the positional relationship between the lane centerline and the lane boundary line. Specifically, for any second discrete point on the lane centerline, this arbitrary second discrete point and its two adjacent discrete points can jointly define a circle, denoted as the second circle. The radius of the second circle is the radius corresponding to this arbitrary second discrete point. It can be understood that if the lane boundary line is considered as a planar curve, the radius of curvature at any second discrete point on this planar curve can also be considered as the radius of that arbitrary second discrete point.

[0065] S420. If the difference between the radius of the second circle and the target radius is less than the first trend threshold, and the distance between the center of the second circle and the center of the target circle is less than the second trend threshold, it is determined that the lane centerline and the lane boundary line tend to be parallel.

[0066] The first and second trend thresholds can also be critical values ​​used to determine whether the lane centerline is smooth. The first and second trend thresholds can be set based on the actual road conditions in the city. They can also be set based on the city level of the road. Furthermore, they can be set based on the geographical features of the area where the lane is located (e.g., hills, plains, basins). Finally, they can be set based on the road level and administrative grade of a particular road. For example, the first trend threshold can be any one of 1m, 0.8m, or 1.2m. The second trend threshold can be any one of 0.1m, 0.2m, or 0.3m.

[0067] Specifically, the radius of the second circle is compared with the target radius. The distance between the center of the second circle and the center of the target circle is determined and compared with a second trend threshold. If the difference between the radius of the second circle and the target radius is less than a first trend threshold, and the distance between the center of the second circle and the center of the target circle is less than the second trend threshold, it indicates that the trend of the lane centerline is the same as the trend of the lane boundary line. Therefore, it can be determined that the lane centerline and the lane boundary line tend to be parallel.

[0068] Based on the positional relationship between the lane centerline and the lane boundary lines, check whether the lane centerline and the lane reference line tend to be parallel, including:

[0069] S430. Based on the parallel positional relationship between the lane centerline and the lane boundary line, determine that the lane centerline and the lane reference line are parallel.

[0070] Specifically, lane boundaries can be used to represent the lane in which a vehicle is located. Therefore, the positional relationship between the lane centerline and the lane reference line can be determined based on the positional relationship between the lane centerline and the lane boundaries. Thus, if the lane centerline and lane boundaries are in a parallel positional relationship, then the lane centerline and lane reference line can be determined to be in a parallel relationship.

[0071] In the above-mentioned lane centerline detection method, when at least two target circles are located on different sides of the lane centerline, the parallel relationship between the lane centerline and the lane boundary line is further utilized to determine that the lane centerline and the lane reference line have the same trend, thereby improving the accuracy of lane centerline detection.

[0072] In some implementations, please refer to Figure 5 Determining the positional relationship between the lane centerline and the lane boundary lines may also include the following steps:

[0073] S510. If the difference between the radius of the second circle and the target radius is not less than the first trend threshold, and / or the distance between the center of the second circle and the center of the target circle is not less than the second trend threshold, it is determined that the lane centerline and the lane boundary line do not tend to be parallel.

[0074] Based on the positional relationship between the lane centerline and the lane boundary lines, check whether the lane centerline and the lane reference line tend to be parallel, including:

[0075] S520. Based on the non-parallel positional relationship between the lane centerline and the lane boundary line, determine that the lane centerline and the lane reference line are not parallel.

[0076] Specifically, the radius of the second circle is compared with the target radius. The distance between the center of the second circle and the center of the target circle is determined and compared with a second trend threshold. If the difference between the radius of the second circle and the target radius is not less than a first trend threshold, or if the distance between the center of the second circle and the center of the target circle is not less than a second trend threshold, it indicates that the trend of the lane centerline is different from that of the lane boundary line. Therefore, it can be determined that the lane centerline and the lane boundary line tend to intersect. Furthermore, based on the positional relationship of the tending intersection between the lane centerline and the lane boundary line, it can be determined that the lane centerline and the lane reference line tend to intersect, that is, the lane centerline is not smooth and needs to be filtered out.

[0077] In the above-mentioned lane centerline detection method, when at least two target circles are located on different sides of the lane centerline, the parallel relationship between the lane centerline and the lane boundary line is further utilized to determine that the lane centerline and the lane reference line have the same trend, thereby improving the accuracy of lane centerline detection.

[0078] In some implementations, please refer to Figure 6 Determining multiple first discrete points on the lane centerline may include the following steps:

[0079] S610. Upsample the lane centerline to obtain multiple initial discrete points in sequence.

[0080] S620. If the distance between any two adjacent initial first discrete points exceeds a preset distance threshold, then add a first discrete point between any two adjacent initial first discrete points.

[0081] S630. Based on multiple initial first discrete points and additional first discrete points, determine multiple first discrete points.

[0082] The preset distance threshold can be set according to the actual situation, such as any one of 3cm, 5cm, or 6cm. Specifically, to discretize the lane centerline, the lane centerline can be upsampled to obtain multiple initial first discrete points in a specific order. The distance between any two adjacent initial first discrete points is detected and compared with the preset distance threshold. If the distance between any two adjacent initial first discrete points exceeds the preset distance threshold, it indicates that the distance between the adjacent initial first discrete points is too large, and a first discrete point needs to be inserted between these two adjacent initial first discrete points. For example, at least one first discrete point can be inserted between any two adjacent initial first discrete points with the preset distance threshold as the interval. The multiple initial first discrete points on the lane centerline, along with the added first discrete points, are used to determine the multiple first discrete points on the lane centerline.

[0083] In the above embodiments, by upsampling the lane centerline and adding a first discrete point between two adjacent initial first discrete points that are far apart, the rationality of the distribution of the first discrete points is improved, providing a data basis for accurately detecting the lane centerline.

[0084] In some implementations, please refer to Figure 7 Obtaining the radii corresponding to multiple first discrete points can include the following steps:

[0085] S710. For any first discrete point, determine two adjacent discrete points that are adjacent to the first discrete point.

[0086] S720. Determine the first circle of any first discrete point based on the position information of any first discrete point and the position information of two adjacent discrete points.

[0087] S730. Take the radius of the first circle at any first discrete point as the radius corresponding to any first discrete point.

[0088] Specifically, multiple first discrete points are distributed along the center line of the lane. Each first discrete point has two adjacent discrete points, such as two adjacent discrete points located on either side of any first discrete point. Since three points can determine a circle, the position information of any first discrete point and the position information of the two adjacent discrete points are used to determine the first circle of any first discrete point, and the radius of the first circle of any first discrete point is taken as the radius corresponding to any first discrete point.

[0089] For example, three consecutive discrete points on the lane centerline; please refer to the following. Figure 1g Let points Y, M, and N have coordinates (x1, y1), (x2, y2), and (x3, y3) respectively. The equation of the circle formed by points Y, M, and N is: x 2+y² + 2gx + 2fy + c = 0. By solving this equation, the center and radius of the circle determined by these three first discrete points can be calculated.

[0090] It is understandable that if the center line of the lane is regarded as a planar curve, the radius of curvature at any first discrete point on the planar curve can also be regarded as the radius of that first discrete point.

[0091] In some implementations, the positional relationship between the lane centerline and the lane reference line is detected based on the comparison results of the radii corresponding to multiple first discrete points and a preset radius threshold, so as to detect whether the lane centerline is smooth. This includes: if there is no target radius smaller than the preset radius threshold among the radii corresponding to multiple first discrete points, it is determined that the lane centerline and the lane reference line tend to be parallel, and then the lane centerline can be determined to be smooth.

[0092] Specifically, by comparing the radius corresponding to the first discrete point with a preset radius threshold, if there is no first discrete point whose radius is less than the preset radius threshold, it can be determined that the lane centerline and the lane reference line tend to be parallel, and thus the lane centerline can be determined to be smooth.

[0093] In the above lane centerline detection method, the positional relationship between the lane centerline and the lane reference line is determined by comparing the radii corresponding to multiple first discrete points with a preset radius threshold, thereby improving the accuracy of smooth lane centerline detection.

[0094] This specification also provides a lane centerline detection method. Please refer to [link / reference]. Figure 8 The lane centerline detection method may include the following steps:

[0095] S802. Upsample the lane centerline to obtain multiple initial discrete points in sequence.

[0096] The lane to which the center line belongs is marked with lane boundaries; each end of the target line segment has a second discrete point.

[0097] S804. If the distance between any two adjacent initial first discrete points exceeds a preset distance threshold, then add a first discrete point between any two adjacent initial first discrete points.

[0098] S806. Based on multiple initial first discrete points and additional first discrete points, determine multiple first discrete points.

[0099] S808. For any first discrete point, determine two adjacent discrete points that are adjacent to the first discrete point.

[0100] S810. Determine the first circle of any first discrete point based on the position information of any first discrete point and the position information of two adjacent discrete points.

[0101] The first circle has a corresponding center and radius.

[0102] S812. Take the radius of the first circle at any first discrete point as the radius corresponding to any first discrete point.

[0103] S814. Compare the radii corresponding to multiple first discrete points with those less than a preset radius threshold.

[0104] S816. If there is no target radius smaller than a preset radius threshold corresponding to multiple first discrete points, and the lane centerline and the lane reference line are determined to be parallel, then the lane centerline can be determined to be smooth.

[0105] Among them, lane reference lines are used to indicate the lane in which the lane center line is located.

[0106] S818. If there is a target radius smaller than a preset radius threshold among the radii corresponding to multiple first discrete points, at least two first target discrete points are determined among the multiple first discrete points on the lane centerline.

[0107] The first target discrete point corresponds to the center of the target circle. Specifically, the first discrete point located at both ends of the lane centerline is determined as the first target discrete point.

[0108] S820. If at least two target circles are located on the same side of the lane centerline, and the lane centerline and lane reference line are determined to be parallel, then the lane centerline can be determined to be smooth.

[0109] S822. If at least two target circles are located on different sides of the lane centerline, the second circle is determined by the second discrete point and two adjacent discrete points adjacent to the second discrete point.

[0110] The second circle has a radius and a center.

[0111] S824. Determine the difference between the radius of the second circle and the target radius. Calculate the distance between the center of the second circle and the center of the target circle.

[0112] S826. If the difference between the radius of the second circle and the target radius is less than the first trend threshold, and the distance between the center of the second circle and the center of the target circle is less than the second trend threshold, it is determined that the lane centerline and the lane boundary line tend to be parallel.

[0113] S828. Based on the parallel positional relationship between the lane centerline and the lane boundary line, if the lane centerline and the lane reference line are determined to be parallel, then the smoothness of the lane centerline can be determined.

[0114] S830. If the difference between the radius of the second circle and the target radius is not less than the first trend threshold, and / or the distance between the center of the second circle and the center of the target circle is not less than the second trend threshold, it is determined that the lane centerline and the lane boundary line do not tend to be parallel.

[0115] S832. Based on the non-parallel positional relationship between the lane centerline and the lane boundary line, it can be determined that the lane centerline and the lane reference line are not parallel, thus indicating that the lane centerline is not smooth.

[0116] This specification provides a lane centerline detection device 900. Please refer to [link / reference]. Figure 9 The lane centerline detection device 900 includes: a discrete point determination module 910, a radius acquisition module 920, and a positional relationship detection module 930.

[0117] The discrete point determination module 910 is used to determine multiple first discrete points on the center line of the lane.

[0118] The radius acquisition module 920 is used to acquire the radius corresponding to the plurality of first discrete points; wherein the radius is the radius of a first circle determined by the first discrete point and two adjacent discrete points adjacent to the first discrete point.

[0119] The positional relationship detection module 930 is used to detect the positional relationship between the lane centerline and the lane reference line based on the comparison result between the radii corresponding to the plurality of first discrete points and a preset radius threshold, so as to detect whether the lane centerline is smooth; wherein, the lane reference line is used to represent the lane in which the lane centerline is located.

[0120] This specification provides a lane line detection device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of the method described in any of the above embodiments.

[0121] This specification provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the methods described in any of the above embodiments. One embodiment of this specification provides a computer program product including instructions that, when executed by a processor of a computer device, enable the computer device to perform the steps of the methods described in any of the above embodiments.

[0122] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0123] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0124] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0125] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0126] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., 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, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0127] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for detecting lane centerlines, characterized in that, The method includes: Determine multiple first discrete points on the centerline of the lane; Obtain the radius corresponding to the plurality of first discrete points; wherein, the radius is the radius of a first circle determined by the first discrete point and two adjacent discrete points adjacent to the first discrete point; Based on the comparison results between the radii corresponding to the plurality of first discrete points and a preset radius threshold, the positional relationship between the lane centerline and the lane reference line is detected to detect whether the lane centerline is smooth; wherein, the lane reference line is used to represent the lane in which the lane centerline is located; If there is a target radius smaller than a preset radius threshold among the radii corresponding to the plurality of first discrete points, then at least two target circle centers are determined; If the centers of the at least two target circles are located on the same side of the lane centerline, it is determined that the lane centerline and the lane reference line tend to be parallel.

2. The method according to claim 1, characterized in that, The step of detecting the positional relationship between the lane centerline and the lane reference line based on the comparison results of the radii corresponding to the plurality of first discrete points and a preset radius threshold includes: If there is a target radius smaller than a preset radius threshold among the radii corresponding to the plurality of first discrete points, at least two first target discrete points are determined among the plurality of first discrete points on the lane centerline; wherein, the at least two first target discrete points correspond to at least two target circle centers; Based on the positional relationship between the lane centerline and the center of the at least two target circles, the positional relationship between the lane centerline and the lane reference line is detected.

3. The method according to claim 2, characterized in that, Determining at least two first target discrete points from among a plurality of first discrete points on the lane centerline includes: The first discrete point located at both ends of the center line of the lane is determined as the first target discrete point.

4. The method according to claim 2, characterized in that, The step of detecting the positional relationship between the lane centerline and the lane reference line based on the positional relationship between the lane centerline and the centers of the at least two target circles includes: If the centers of the at least two target circles are located on different sides of the lane centerline, determine the lane boundary corresponding to the lane to which the lane centerline belongs; Based on the positional relationship between the lane centerline and the lane boundary line, detect whether the lane centerline and the lane reference line tend to be parallel.

5. The method according to claim 4, characterized in that, The lane boundary line includes a target line segment corresponding to the position of the lane center line, and each end of the target line segment has a second discrete point; before detecting whether the lane center line and the lane reference line tend to be parallel based on the positional relationship between the lane center line and the lane boundary line, the method further includes: The second circle is determined by the second discrete point and two adjacent discrete points adjacent to the second discrete point. If the difference between the radius of the second circle and the target radius is less than the first trend threshold, and the distance between the center of the second circle and the center of the target circle is less than the second trend threshold, it is determined that the lane centerline and the lane boundary line tend to be parallel. The step of detecting whether the lane centerline and the lane reference line tend to be parallel based on the positional relationship between the lane centerline and the lane boundary line includes: Based on the parallel positional relationship between the lane centerline and the lane boundary line, it is determined that the lane centerline and the lane reference line are approximately parallel.

6. The method according to claim 5, characterized in that, Before detecting whether the lane centerline and the lane reference line tend to be parallel based on the positional relationship between the lane centerline and the lane boundary line, the method further includes: If the difference between the radius of the second circle and the target radius is not less than the first trend threshold, and / or the distance between the center of the second circle and the center of the target circle is not less than the second trend threshold, it is determined that the lane centerline and the lane boundary line do not tend to be parallel. The step of detecting whether the lane centerline and the lane reference line tend to be parallel based on the positional relationship between the lane centerline and the lane boundary line includes: Based on the non-parallel positional relationship between the lane centerline and the lane boundary line, it is determined that the lane centerline and the lane reference line are not parallel.

7. The method according to claim 1, characterized in that, Determining the plurality of first discrete points on the centerline of the lane includes: The lane centerline is upsampled to obtain multiple initial discrete points in a specific order; If the distance between any two adjacent initial first discrete points exceeds a preset distance threshold, then an additional first discrete point is added between the two adjacent initial first discrete points. The plurality of first discrete points are determined based on the plurality of initial first discrete points and the added first discrete points.

8. The method according to claim 1, characterized in that, Obtaining the radii corresponding to the plurality of first discrete points includes: For any first discrete point, determine two adjacent discrete points that are adjacent to the first discrete point; Based on the position information of the arbitrary first discrete point and the position information of two adjacent discrete points adjacent to the arbitrary first discrete point, the first circle of the arbitrary first discrete point is determined. The radius of the circle at any first discrete point is taken as the radius corresponding to the arbitrary first discrete point.

9. The method according to claim 1, characterized in that, The step of detecting the positional relationship between the lane centerline and the lane reference line based on the comparison results of the radii corresponding to the plurality of first discrete points with a preset radius threshold includes: If none of the radii corresponding to the plurality of first discrete points is smaller than a preset radius threshold, it is determined that the lane centerline and the lane reference line tend to be parallel.

10. A lane centerline detection device, characterized in that, The apparatus for implementing the method according to any one of claims 1 to 9, comprising: A discrete point determination module is used to determine multiple first discrete points on the center line of the lane; A radius acquisition module is used to acquire the radius corresponding to the plurality of first discrete points; wherein, the radius is the radius of a first circle determined by the first discrete point and two adjacent discrete points adjacent to the first discrete point; The positional relationship detection module is used to detect the positional relationship between the lane centerline and the lane reference line based on the comparison results of the radii corresponding to the plurality of first discrete points and a preset radius threshold, so as to detect whether the lane centerline is smooth; wherein, the lane reference line is used to represent the lane in which the lane centerline is located.

11. A lane line detection device, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 9.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 9.

Images