A method and system for accurately determining a road centerline driven by a laser point cloud

Abstract

This invention relates to a method and system for accurate road centerline determination driven by laser point clouds. The method is based on laser point clouds, using a circular method to extract the road contour from the road surface point cloud and further using a Venn diagram to extract the road centerline; it then uses a threshold judgment to determine the reflection intensity of the road surface point cloud to extract lane markings; based on the obtained road centerline, it establishes a lane marking data index, dividing the extracted lane markings into inner and outer lane markings; it uses the nearest road center point as the index and calculates the average value of the inner and outer lane markings to obtain the lane centerline; finally, it calculates and corrects the deviation between the extracted lane centerline and the road centerline to determine the road centerline. Compared with existing technologies, this invention reduces the impact of vehicle trajectory fluctuations and long-distance laser acquisition, and can accurately determine the road centerline from laser point clouds, offering advantages such as good adaptability and high accuracy.

Description

Technical Field

[0001] This invention belongs to the field of vehicle-mounted laser point cloud data processing, and specifically relates to a method and system for accurately determining the centerline of a road driven by laser point cloud. Background Technology

[0002] The road centerline generally refers to the center line of the road width. However, for roads with one-way traffic or those that have a gradual widening transition, the centerline may not be the actual road centerline, but rather the center line of a lane in the same direction. Taking a widened section of the road as an example, when the radius of the horizontal curve is equal to or less than 250m, the road should be widened inside the horizontal curve. In this case, the road centerline will deviate from the actual road centerline (lane centerline).

[0003] Current technologies typically determine road centerlines using vehicle-mounted laser systems to track vehicle movements. However, this method is susceptible to errors when vehicles deviate from their tracks due to complex road conditions, leading to inaccuracies in the extracted centerline. Furthermore, this approach is unsuitable for situations where trajectory data is missing (stationary laser scanning). To improve the algorithm's versatility, some researchers use lane markings to obtain the road centerline. However, the accuracy of the extracted lane markings is compromised when laser acquisition is performed at long distances or when the markings are worn. Other researchers extract the road centerline from the road contour for subsequent information extraction, but the road centerline cannot accurately describe the changes in road sections (widening transition zones). Summary of the Invention

[0004] The technical problem to be solved by this invention is to propose a method and system for accurate determination of road centerline driven by laser point cloud. The method uses the circular method to extract the road contour, uses the Venn diagram to extract the road centerline, extracts the lane markings and divides them into inner and outer lane markings, calculates the average value of the inner and outer lane markings, obtains the lane centerline, calculates and corrects the deviation between the extracted lane centerline and the road centerline, and determines the road centerline. This reduces the impact of driving trajectory fluctuations and long-distance laser acquisition.

[0005] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0006] A method for accurately determining the centerline of a road driven by laser point cloud includes the following steps:

[0007] S1. Use the RANSAC algorithm to extract road surface point cloud from laser point cloud.

[0008] S2. Use the circular rolling method to extract the road contour from the road surface point cloud.

[0009] S3. Use the Venn diagram to extract the road centerline from the road profile.

[0010] S4. Threshold judgment is performed on the reflection intensity of the road surface point cloud to extract lane marking points.

[0011] S5. Based on the obtained road centerline, establish a lane marking data index and divide the extracted lane markings into inner and outer lane marking point clusters.

[0012] S6. For lane marking points clustered together on the inner and outer sides, index them using the same road center point, calculate the average three-dimensional coordinates of the inner and outer lane markings, and obtain the lane centerline.

[0013] S7. Calculate and correct the deviation between the lane centerline and the road centerline to determine the road centerline.

[0014] Furthermore, in step S1, the specific content of extracting the road surface point cloud is as follows: when d i <d t If , it means that the point lies on the plane, where d i Let d be the distance from any point i in the laser point cloud to the plane. t The distance threshold is used to count the number of points on the plane. The number of points on the plane is set to rn. The point cloud plane with a number of points on the plane ≥ rn is the road surface point cloud.

[0015] Furthermore, in step S4, the specific content of extracting lane marking points is as follows: when I j ≥I H This point is the lane marking point; where I j Let I be the reflection intensity of the j-th point in the road surface point cloud. H This is the threshold value for the reflection intensity of the caliper.

[0016] Furthermore, in step S5, the specific details of the inner and outer lane markings are as follows:

[0017] For any point p in the lane markings, use a Kd-tree data structure to search for the nearest road center point C from the road centerline. i and the corresponding shortest distance d min , denoted as (C i ,d min );

[0018] If the x-coordinate of point p is greater than the x-coordinate of the road center point, then d min A positive sign indicates that the marking point is outside the road centerline; conversely, a negative sign indicates that d... min A negative sign indicates that the marking point is inside the center line of the road.

[0019] Furthermore, in step S6, the specific details of obtaining the lane centerline are as follows:

[0020] Index C using any road center point i Extract the corresponding inner and outer lane marking points (X) from the cloud of inner and outer lane marking points respectively. li Y li Zli ) and (X ri Y ri Z ri Calculate the average three-dimensional coordinates of the inner and outer lane markings, and use this average as the coordinates of the lane centerline. The specific formula is as follows:

[0021]

[0022]

[0023]

[0024] Repeat the traversal until the coordinates of all points in the lane markings are calculated to obtain the lane centerline.

[0025] Furthermore, in step S7, the specific steps for determining the road centerline are as follows:

[0026] S701, Based on any road center point C i coordinates (X) ci Y ci Z ci ) and the corresponding lane centerline T i coordinates (X) mi Y mi Z mi Project the two points onto the XY plane and calculate the deviation dL. i and T i tangent angle θ at the point i The specific formula is as follows:

[0027]

[0028] θ i =arctan(k i )

[0029] In the formula, k i The curve formed by the lane centerline in the XY plane at T i The slope at that point.

[0030] S702, Based on the deviation dL i and T i tangent angle θ at the point i X of the road centerline ci Y ci The coordinates are corrected to obtain the X-axis of the road centerline. i '、Y i 'Coordinates, the specific formula is:'

[0031] X′ i =X ci -dL i ×cos(θi )

[0032] Y′ i =Y ci +dL i ×sin(θ i ).

[0033] S703, based on the X of the road centerline i '、Y i 'Coordinates, the Z-axis corresponding to this location is obtained by searching in the road surface point cloud.' i The coordinates, i.e., the coordinates of the road centerline, are (X... i ', Y i ', Z i ').

[0034] Furthermore, this invention also proposes a laser point cloud-driven system for precise determination of road centerlines, including...

[0035] The road surface point cloud extraction module is used to extract road surface point clouds from laser point clouds using the RANSAC algorithm.

[0036] The road contour extraction module is used to extract road contours from road surface point clouds using the circular method.

[0037] The road centerline extraction module is used to extract the road centerline from the road profile using a Venn diagram.

[0038] The lane marking point extraction module is used to determine the reflection intensity of the road surface point cloud by threshold and extract lane marking points.

[0039] The inner and outer lane marking point set division module is used to establish a lane marking data index based on the obtained road centerline and divide the extracted lane markings into inner and outer lane marking point sets.

[0040] The lane centerline acquisition module is used to aggregate the inner and outer lane marking points, index them with the same road center point, calculate the average three-dimensional coordinates of the inner and outer lane markings, and obtain the lane centerline.

[0041] The road centerline determination module is used to calculate and correct the deviation between the lane centerline in the lane centerline module and the road centerline to determine the road centerline.

[0042] Furthermore, the present invention also proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of a laser point cloud-driven method for accurately determining the road centerline.

[0043] Furthermore, the present invention also proposes a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the laser point cloud-driven method for precise determination of the road centerline.

[0044] The present invention adopts the above technical solution, and its significant technical effects compared with the prior art are as follows:

[0045] (1) Based on laser point cloud, this invention can accurately extract the centerline of the road, which is conducive to carrying out geometric line calculation and realizing road digitization.

[0046] (2) The present invention constructs a data index for lane markings based on the distance between the extracted road centerline and the corresponding center point. The relative position of the lane markings on the road can be determined according to the three-dimensional coordinates of the lane markings, so as to distinguish between inner and outer lane markings.

[0047] (3) The present invention merges and corrects the lane centerline and the road centerline to determine the road centerline, avoiding the point cloud noise data and deviations in widened road sections caused by the use of a single lane marking or road centerline in conventional methods.

[0048] (4) This invention avoids the need to correct the Z coordinate in the YZ or XZ plane, thus reducing computational complexity. At the same time, the Z coordinate is obtained from the existing point cloud based on the X and Y coordinates, ensuring the rationality and accuracy of the data.

[0049] (5) This invention only uses basic data of laser point cloud, and has high applicability. Attached Figure Description

[0050] Figure 1 This is a flowchart illustrating the overall implementation of the present invention.

[0051] Figure 2 This is a schematic diagram of the road surface point cloud extraction for a certain part of the road in this invention.

[0052] Figure 3 This is a schematic diagram of the extraction of the center line of a certain part of the road in this invention.

[0053] Figure 4 This is a schematic diagram illustrating the calculation and correction of the deviation between the lane centerline and the center line in this invention.

[0054] Figure 5 This is a laser point cloud of a portion of the ramps in an embodiment of the present invention.

[0055] Figure 6 This is the result of road centerline extraction and comparison in an embodiment of the present invention. Detailed Implementation

[0056] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0057] This embodiment describes a method for accurately determining the centerline of a road driven by laser point cloud. Figure 1 As shown, it includes the following steps:

[0058] S1. The `segment_plane` function from the Open3d library in Python is used to extract road surface point clouds from laser point clouds. The main parameters are the distance threshold `distance_threshold`, the number of plane points `ransac_n`, and the number of iterations `num_iterations`. Specifically, the distance from any point `i` in the laser point cloud to the plane is `d`. i Set the distance threshold to d t When d i <d t If the number of points on the plane is 1, it means that the point lies on the plane. Count the number of points on the plane, and set the number of points on the plane as rn. The point cloud plane with a number of points on the plane ≥ rn is the road surface point cloud. Figure 2 As shown.

[0059] S2. Use the circular rolling method to extract the road contour from the road surface point cloud.

[0060] The contour was extracted using the `boundary` function in Matlab. A circle of radius R was rolled around the road surface point cloud. Two arbitrary points, p1 and p2, were selected within the point set. A circle of radius R was drawn through these two points. If no other points were inside the circle, p1p2 could be considered the boundary contour. When the distance between points p1 and p2 was less than 2R, two circles of radius R would pass through these two points. If at least one of these circles contained no other points, p1p2 was the boundary line, because the area outside the boundary line was empty, and the outer circle would not contain any other points.

[0061] S3. Use the `midlines_clean` function from the `midlines` library in R software to extract the road centerline from the road outline, such as... Figure 3 As shown.

[0062] S4. Threshold the reflection intensity of the road surface point cloud and extract lane marking points. Specifically, the reflection intensity of the j-th point in the road surface point cloud is I. j The reflection intensity threshold I of the gait H ; when I j ≥I H This point is the lane marking point. Among them, I... H =125.

[0063] S5. Based on the obtained road centerline, establish a lane marking data index, and divide the extracted lane markings into inner and outer lane marking point agglomerations. The specific content is as follows:

[0064] For any point p in the lane markings, its three-dimensional coordinates are (522219.959, 3638034.298, 37.86). Use a Kd-tree data structure to search for the nearest road center point C from the road centerline. i The value is 1322, and the corresponding shortest distance d. min The value is 0.9819, denoted as (1322, 0.9819).

[0065] Since the x-coordinate of point p is greater than the x-coordinate of the road center point, then d min Take the positive sign, and place the marking point outside the center line of the road.

[0066] S6. For lane marking points clustered together on both the inner and outer sides, index them using the same road center point, calculate the average three-dimensional coordinates of the inner and outer lane markings, and obtain the lane centerline. The specific steps are as follows:

[0067] Index C using any road center point i Its three-dimensional coordinates are (522250.080, 3638023.899, 38.747). From the inner and outer lane marking point clouds, the corresponding inner and outer lane marking points (522226.0931, 3638041.722, 37.83) and (522274.4899, 3638006.699, 39.586) are extracted respectively. The average three-dimensional coordinates of the inner and outer lane markings are calculated as the coordinates of the lane centerline. The specific formula is as follows:

[0068]

[0069]

[0070]

[0071] Repeat the traversal until the coordinates of all points in the lane markings are calculated to obtain the lane centerline.

[0072] Calculations show that X mi =522250.2915, Y mi =3638024.211, Y mi =38.708.

[0073] S7. Calculate and correct the deviation between the lane centerline and the road centerline to determine the road centerline. The specific steps are as follows:

[0074] S701, Based on any road center point Ci The coordinates (522250.080, 3638023.899, 38.747) and the corresponding lane centerline T i Given the coordinates (522250.2915, 3638024.211, 38.708), project the two points onto the XY plane and calculate the deviation dL. i and T i tangent angle θ at the point i ,like Figure 4 As shown, the specific formula is:

[0075]

[0076] θ i =arctan(k i )

[0077] In the formula, k i The curve formed by the lane centerline in the XY plane at T i The slope at that point.

[0078] The distance deviation dL was calculated. i =0.3765, k=1.4728, θ i It is 0.9743.

[0079] S702, Based on the deviation dL i and T i tangent angle θ at the point i X of the road centerline ci Y ci The coordinates are corrected to obtain the X-axis of the road centerline. i '、Y i 'Coordinates, the specific formula is:'

[0080] X′ i =X ci -dL i ×cos(θ i )

[0081] Y′ i =Y ci +dL i ×sin(θ i ).

[0082] S703, based on the X of the road centerline i '、Y i 'Coordinates, the Z-axis corresponding to this location is obtained by searching in the road surface point cloud.' i The coordinates, i.e. the coordinates of the road centerline, are (522249.8685, 3638024.211, 38.748).

[0083] The experimental data used in this embodiment is point cloud data of the ramps of an interchange in Huai'an City, such as... Figure 5 As shown. Figure 6 As shown, comparing the corrected road centerline with the road centerline in the CAD drawing reveals that the corrected road centerline is basically consistent with the CAD drawing, demonstrating good application results.

[0084] This invention also proposes a laser point cloud-driven system for precise determination of road centerlines, including a road surface point cloud extraction module, a road contour extraction module, a road centerline extraction module, a lane marking point extraction module, an inner and outer lane marking point cloud set division module, a lane centerline acquisition module, a road centerline determination module, and a computer program that can run on a processor. It should be noted that each module in the above system corresponds to a specific step of the method provided in this invention embodiment, possessing the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the method provided in this invention embodiment.

[0085] This invention also proposes an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. It should be noted that each module in the above system corresponds to a specific step of the method provided in this invention, possessing the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the method provided in this invention.

[0086] This invention also proposes a computer-readable storage medium storing a computer program. It should be noted that each module in the above system corresponds to a specific step of the method provided in this invention, possessing the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the method provided in this invention.

[0087] The preferred embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make numerous modifications and variations based on the concept of the present invention without creative effort. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning, or limited experimentation on the basis of existing technology should be within the scope of protection defined by the claims.

Claims

1. A method for accurate determination of a road centerline driven by a laser point cloud, characterized in that, Comprise the following steps: S1, extracting road surface point cloud from laser point cloud by using RANSAC algorithm; S2, extracting road profile from road surface point cloud by using rolling method; S3, extracting road center line from road profile by using Vino graph; S4, threshold judgment is carried out on the reflection intensity of road surface point cloud, and lane marking point is extracted; S5, according to the obtained road center line, lane marking data index is established, and the extracted lane marking is divided into inside and outside lane marking point cloud set; S6, the inside and outside lane marking point cloud set is indexed with the same road center point, the three-dimensional coordinate average value of the inside and outside lane marking is calculated, and the lane center line is obtained; S7, the deviation of lane center line and road center line is calculated and corrected to determine the road center line; Specifically: S701, according to the coordinates (X i , Y ci , Z ci ) of any road center point C ci and the coordinates (X i , Y mi , Z mi ) of the corresponding lane center line T mi , project the two points to the XY plane, calculate the deviation dL i and the tangent angle θ i at T i , the specific formula is: In the formula, k i is the slope of the curve formed by the lane centerline in the XY plane at T i ; S702、According to the deviation dL i and T i The tangent angle θ i , the X ci , Y ci coordinate of the road center line is corrected to obtain the X i ', Y i ' coordinate of the road center line, and the specific formula is: X i ′ = X ci - dL i x cos(θ i ) Y i ′ = Y ci + dL i x sin(θ i ); S703、According to the X i ’ coordinate of the road centerline, search for the corresponding Z i ’ coordinate in the road surface point cloud. i ’ coordinate, that is, the coordinate of the road centerline is (X i ’, Y i ’, Z i ’ ).

2. The method of claim 1, wherein, In step S1, the specific content of extracting the road surface point cloud is: when d i <d t represents that the point is on the plane, where d i is the distance of any point i in the laser point cloud to the plane, d t is the distance threshold; the number of points on the plane is counted, and the number of points on the plane is set as rn. When the number of points on the plane ≥ rn, the point cloud plane is the road surface point cloud.

3. The method of claim 1, wherein, In step S4, the specific content of extracting lane marking points is: when I j ≥I H , the point is a lane marking point; wherein I j is the reflection intensity of the jth point in the road point cloud, and I H is a lane marking reflection intensity threshold.

4. The method of claim 1, wherein, In step S5, the specific content of inside and outside lane marking division is: For any point p in the lane marking, search for the nearest road center point C from the road centerline using Kd-tree data structure i and the corresponding shortest distance d min , denoted as (C i ,d min ) If the abscissa of the p point is greater than the abscissa of the road center point, then d min is taken as positive, indicating that the reticle point is outside the road center line; otherwise, d min is taken as negative, indicating that the reticle point is inside the road center line.

5. The method of claim 4, wherein, In step S6, the specific content of obtaining lane center line is: Indexing C with any road center point i , respectively, from the inner and outer lane marking point cloud sets, the corresponding inner and outer lane marking points (X li , Y li , Z li ) and (X ri , Y ri , Z ri ) are extracted, and the three-dimensional coordinate average values of the inner and outer lane markings are calculated as the coordinates of the lane center line, and the specific formula is as follows: Repeat traversal until the coordinate calculation of all points in lane marking is completed, and the lane center line is obtained. 6.A system for accurate determination of road centerline driven by laser point cloud, characterized in that, Comprise Road surface point cloud extraction module, for extracting road surface point cloud from laser point cloud by using RANSAC algorithm; Road profile extraction module, for extracting road profile from road surface point cloud by using rolling method; Road center line extraction module, for extracting road center line from road profile by using Vino graph; Lane marking point extraction module, for threshold judgment on the reflection intensity of road surface point cloud, and extracting lane marking point; Inside and outside lane marking point cloud set division module, for establishing lane marking data index according to the obtained road center line, and dividing the extracted lane marking into inside and outside lane marking point cloud set; Lane center line acquisition module, for indexing the inside and outside lane marking point cloud set with the same road center point, calculating the three-dimensional coordinate average value of the inside and outside lane marking, and obtaining the lane center line; Road center line determination module, for calculating the deviation of lane center line in lane center line module and road center line and correcting to determine the road center line; Specifically: Step 1, according to the coordinates (X ci , Y ci , Z ci ) of any road center point C i and the coordinates (X mi , Y mi , Z mi ) of the corresponding lane center line T i , project the two points to the XY plane, calculate the deviation dL i and the tangent angle θ i at T i , the specific formula is: θ i = arctan(k i ); In the formula, k i is the slope of the curve formed by the lane centerline in the XY plane at T i ; Step 2: Based on the deviation dL i and T i tangent angle θ at the point i X of the road centerline ci Y ci The coordinates are corrected to obtain the X-axis of the road centerline. i '、Y i 'Coordinates, the specific formula is:' X i ′= X ci -dL i × cos(θ i ) Y i ′ = Y ci + dL i x sin(θ i ); Step 3, according to the X i ’ coordinate of the road centerline, search for the corresponding Z i ’ coordinate in the road surface point cloud, that is, the coordinates of the road centerline are (X i ’, Y i ’, Z i ’). i ’) 7. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to realize the steps of the method in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, the computer-readable storage medium being characterized by, The computer program is run by the processor to execute the method in any one of claims 1 to 5.

Images