Expressway stake number positioning method and device based on point cloud data element extraction

Abstract

The invention provides an expressway stake mark positioning method and equipment based on point cloud data element extraction, and the method comprises the steps: carrying out the preprocessing of original point cloud data through employing a slope method ground filtering and statistical outlier elimination filtering mode, and extracting a complete road surface; on the basis of the extracted complete road surface data, performing point cloud alignment, slicing and extracting road edges, and preliminarily solving a road center point by utilizing the geographic coordinate center of each point of the extracted road edges; skeleton points are randomly initialized, the positions of the skeleton points are iteratively optimized, then the skeleton points are subjected to sparse processing, and fitting is achieved through minimum spanning tree connection to generate a continuous center line of a road; and matching and positioning kilometer stake points to realize accurate assignment of stake numbers. According to the method, the expressway elements can be efficiently and accurately extracted from massive point cloud data and converted into stake number positioning information, the defects of a traditional method in efficiency and positioning precision are overcome, and the accuracy and working efficiency of expressway positioning are remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent processing based on point cloud data, and designs a method and device for highway mileage positioning based on element extraction from point cloud data. Background Art

[0002] As an important part of modern transportation, the safe and efficient operation of highways depends on the accurate grasp of road information. In the processes of highway maintenance, traffic incident handling, emergency rescue, etc., the demand for precise positioning of target location points is increasing day by day. Traditional highway positioning methods mainly rely on on-site manual surveys or mapping methods based on two-dimensional map data. These methods are not only time-consuming and laborious, but also limited by problems such as manual operation accuracy, environmental conditions, and data update lag, and it is difficult to meet the requirements of modern highway management for efficiency and accuracy.

[0003] With the development of three-dimensional scanning technologies such as Light Detection and Ranging (LiDAR), it has become possible to automatically extract and position highway elements using point cloud data. Point cloud data can provide rich three-dimensional spatial information, including the geometric shape of the road, edge positions, guardrails, signs, bridges, tunnels and other structures, providing new ideas for the precise modeling, change detection and intelligent traffic management of highways. However, in practical applications, there are still the following challenges and technical difficulties in point cloud data for highway element extraction and mileage positioning:

[0004] 1. Complex scenes affect the accuracy of element extraction

[0005] The highway environment is often complex, and there may be various structures such as viaducts, slope changes, route intersections, guardrails, traffic signs, street lights, etc. along the line. At the same time, factors such as vegetation occlusion, vehicle interference, and reflection characteristics of road surfaces with different materials may cause noise, data discontinuity or uneven point density in the point cloud data, thereby affecting the precise extraction of road surfaces, road boundaries and auxiliary facilities. How to improve the quality of point cloud data in complex scenes and effectively separate key highway elements is a key technical challenge.

[0006] 2. Difficulties in coordinate transformation and mileage transformation of point cloud data

[0007] The mileage (mileage identifier) of a highway is an important reference information for road management and navigation, but point cloud data usually uses geographical coordinates (such as WGS84 or UTM) and cannot be directly mapped to the highway mileage system. Mileage is usually cumulative mileage information defined along the center line of the road, while point cloud data is a discrete set of three-dimensional points in space, and there are great challenges in direct matching. Traditional coordinate transformation methods require high-precision road center line data and perform complex projection transformations and distance calculations. Therefore, improving the automation and accuracy of mileage matching is an urgent problem to be solved.

[0008] 3. Computational Efficiency Issues in Large-Scale Point Cloud Data Processing

[0009] Due to the long length of highways, the amount of point cloud data obtained is extremely large, usually containing millions or even billions of points. Directly processing the full amount of point cloud data not only involves huge computational costs, low storage and retrieval efficiency, but also is easily limited by computing resources. Therefore, efficient data organization, block processing, point cloud downsampling, and parallel computing strategies are required to improve the speed of data processing and storage management capabilities.

[0010] In summary, the existing methods for processing highway point cloud data still have deficiencies in terms of accuracy, computational efficiency, and station number mapping. Therefore, there is an urgent need to develop an efficient and accurate method for processing point cloud data to achieve automatic extraction of highway elements and construct a high-precision mapping system from point cloud coordinates to station numbers, so as to improve the application capabilities of highway information management and intelligent transportation systems. Summary of the Invention

[0011] Aiming at the deficiencies in extracting the road surface and road edges from existing point cloud highway data, as well as the inconvenience of coordinate systems such as longitude and latitude of point cloud data in highway positioning, the purpose of the present invention is to provide a highway station number positioning method and device based on point cloud data element extraction.

[0012] The technical solution provided by the present invention is a highway station number positioning method based on point cloud data element extraction, which performs the following processing:

[0013] Preprocess the original point cloud data using the slope method ground filtering and statistical outlier removal filtering methods to extract the complete road surface;

[0014] Based on the extracted complete road surface data, perform point cloud alignment and then slice to extract the road edges, and preliminarily calculate the road center point using the geographical coordinate center of each point on the extracted road edges;

[0015] Randomly initialize the skeleton points, iteratively optimize the positions of the skeleton points, then sparsify the skeleton points, and connect them through the minimum spanning tree to achieve fitting and generating the continuous center line of the road;

[0016] Perform the matching and positioning of kilometer post points to achieve accurate assignment of station numbers.

[0017] Moreover, when extracting the complete road surface, the slope method ground filtering calculates the angle between the normal vector of each point and the vertical direction to screen out the ground points; the statistical outlier removal filtering method removes the abnormal points in the point cloud by setting the neighborhood point number threshold, thereby extracting the complete road surface.

[0018] Moreover, when aligning the point cloud, the road direction is made parallel to the slicing direction through rotational transformation.

[0019] Moreover, the continuous centerline of the road is generated by fitting, and the implementation method includes randomly selecting initial skeleton points from the road point cloud data, iteratively updating the positions of each skeleton point, calculating the attribution of each point to the skeleton point using the L1 distance, and updating the skeleton point with the neighborhood median; removing redundant skeleton points that are too close through sparsification processing, and connecting the skeleton points using the minimum spanning tree algorithm; fitting the complete centerline using a B-spline curve.

[0020] Moreover, when fitting the road centerline using a B-spline curve, a multi-segment fitting strategy is adopted to adapt to the complex changes in the road shape, and the fitting accuracy is optimized by the least squares method to ensure the smoothness and continuity of the centerline.

[0021] Moreover, when matching and positioning the kilometer post points, an error correction mechanism is adopted. After positioning every 1000 points, the error is corrected by matching the actual kilometer post position to ensure the accuracy and reliability of the kilometer post number positioning.

[0022] Moreover, the longitude and latitude information of the integer kilometer posts of a section of highway is input, and all specific positions of this section are retrieved with a precision reaching the meter level.

[0023] On the other hand, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the above-mentioned highway kilometer post positioning method based on point cloud data element extraction is implemented.

[0024] On the other hand, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned highway kilometer post positioning method based on point cloud data element extraction is implemented.

[0025] On the other hand, the present invention also provides a computer program product, including a computer program. When the computer program is executed by a processor, the above-mentioned highway kilometer post positioning method based on point cloud data element extraction is implemented.

[0026] Through the above technical solutions, the present invention can effectively and quickly extract highway elements from large-density point cloud data of multi-frame fusion in a loop, directly fit the centerline of a highway with a length of more than 100 km using the proposed L1-centraline algorithm, convert it into the positioning information of highway kilometer posts, and maintain the accuracy at the level of more than 1 meter, solve the problems existing in the prior art, improve the accuracy and efficiency of highway positioning, and have important application value and broad application prospects. Description of the Drawings

[0027] Figure 1 This is a schematic flowchart of the method for extracting highway elements and locating the mileage based on point cloud data according to an embodiment of the present invention.

[0028] Figure 2 This is an effect diagram of extracting the highway road surface based on point cloud data according to an embodiment of the present invention.

[0029] Figure 3 This is an effect diagram of extracting the road edge points and the road center line of the highway based on point cloud data according to an embodiment of the present invention.

[0030] Figure 4 This is an effect diagram of extracting the 123 km highway road data based on point cloud data according to an embodiment of the present invention.

[0031] Figure 5 This is a comparison of the results of the road center line fitted by the L1 - centraline algorithm and the L1 point cloud backbone extraction algorithm based on point cloud data according to an embodiment of the present invention.

[0032] Figure 6 This is a diagram of the positioning method for converting to the highway mileage based on point cloud data according to an embodiment of the present invention. Specific embodiments

[0033] The following specifically describes the technical solution of the present invention in conjunction with the drawings and embodiments.

[0034] The present invention proposes a new point cloud data processing method for highway element extraction and mileage positioning. This method preferably combines and adopts technical means such as multi - stage filtering, point cloud alignment slicing, B - spline curve fitting, and cyclic positioning and correction to achieve accurate extraction of highway elements and accurate positioning of the mileage; at the same time, a new point cloud data processing flow is designed, which can better extract element information such as the road surface, edge, and center line of the highway and achieve efficient matching with the mileage.

[0035] Embodiment 1

[0036] Refer to Figures 1 - 5 , the present invention provides a method for extracting highway elements and locating the mileage based on point cloud data, including the following steps:

[0037] Step a, using the slope - based ground filtering and statistical outlier removal filtering method to extract the complete road surface;

[0038] The present invention proposes to use the slope - based ground filtering to screen out the ground points by calculating the angle between the normal vector of each point and the vertical direction; the statistical outlier removal filtering removes the abnormal points in the point cloud by setting the threshold of the number of neighborhood points, thereby extracting the complete road surface.

[0039] Step b: Align the extracted complete and accurate road surface data, slice it after point cloud alignment to extract the road edge. During the point cloud alignment process, the point cloud needs to be rotated to be parallel to the slice direction to ensure the accuracy of edge extraction; preliminarily obtain the road center point using the geographical coordinate center of each point on the road edge extracted by road extraction.

[0040] Step c: Randomly initialize the skeleton points, iteratively optimize the positions of the skeleton points, then sparsify the skeleton points, and connect them through the minimum spanning tree to fit and generate the continuous centerline of the road;

[0041] Due to the large deviation points in the center points obtained in step b caused by vehicle blocking and tree interference, etc., it will have a huge impact on the fitting of the complete curve. To solve this problem, the L1-centraline algorithm is proposed to extract the complete road center backbone. The implementation method includes first randomly selecting a certain number of points from the road point cloud data as the initial skeleton points, avoiding the bias of the point cloud distribution. Then, by iteratively updating the position of each skeleton point, calculate the attribution of each point to the skeleton point using the L1 (Manhattan) distance, and update the skeleton point with the neighborhood median. Next, set the L1 distance to 1 meter for sparsification processing to remove redundant skeleton points that are too close, making the skeleton structure more concise. Finally, use the minimum spanning tree (MST) algorithm to connect the skeleton points to ensure that the connection path of the skeleton is the shortest and conforms to the actual structure. Finally, use the B-spline curve to fit the complete centerline;

[0042] When using the B-spline curve to fit the road centerline, it is preferably recommended to adopt a multi-segment fitting strategy to adapt to the complex changes in the road shape, and optimize the fitting accuracy through the least squares method to ensure the smoothness and continuity of the centerline.

[0043] Step d: Match and locate the kilometer post points to achieve accurate assignment of the mileage number:

[0044] First, select a point that completely matches the kilometer post. Specifically, the line connecting this point and the post completely intersects the tangent of the road. Then, increase the mileage number point by point along the road forward. Thus, all points on the road centerline are given specific mileage number values. To avoid the accumulation of errors, after every 1000 points of positioning, that is, when the route advances 1 km, it is necessary to judge whether there is a large deviation between the point cloud and the next integer kilometer post. If this occurs, one of the points needs to be selected for complete matching, and then the positioning is looped to ensure the accuracy and reliability of the positioning.

[0045] Example 2

[0046] Based on the process of Example 1, further in step a, the following method is used to extract the complete road surface:

[0047] Use the slope method for ground filtering to select points that conform to the road surface slope as road surface points. During the calculation of point cloud normals, it is assumed that each point has a normal, which represents the surface orientation at the location of the point. The normal is calculated through the geometric features of neighboring points in the point cloud, and normal estimation can be performed through local plane fitting. The calculation of the normal takes into account the points within the neighborhood, and the spatial distribution of these points affects the direction of the normal. During the process of calculating the angle between the normal and the horizontal plane, first obtain the ground points in advance. The z-component of the normal can be used to calculate the angle between the point and the horizontal plane. Assuming the vector of the normal is (nx, ny, nz) and the normal of the horizontal plane is (0, 0, 1), the angle θ between the normal and the horizontal plane can be calculated by the following formula:

[0048]

[0049] where θ is the angle between the normal and the horizontal plane, nx is the X-axis component of the normal, ny is the Y-axis component of the normal, and nz is the Z-axis component of the normal.

[0050] Use the statistical outlier removal filtering method to remove the miscellaneous points that cannot be removed by the slope method. Specifically, for each point, calculate the distances from all points within its neighborhood to this point, and calculate the average value mean dist and the standard deviation std dist Remove the points that are greater than the preset threshold range. The calculation formula is as follows:

[0051]

[0052] where d i is the distance between point i and the current point, and N is the number of points within the neighborhood.

[0053] Since on first-class highways and expressways, to ensure the normal driving of vehicles on the main line lanes, the maximum slope specified is 3% (4 - 5% in mountainous and hilly areas), and the section where the embodiment is applied belongs to the plain area, the slope threshold is set to 3%. To balance the accuracy of miscellaneous point removal effect and the integrity of the road surface, the selected threshold is: d i is 2 meters, and N is 20. During specific implementation, the corresponding threshold parameters can be adjusted according to needs.

[0054] For the effect of extracting the road surface point cloud in the embodiment, refer to the appendix Figure 2 , because the point cloud data in this application scenario is the data after multi-frame fusion with a large data density, the present invention preferably uses the above simple calculation process to quickly extract the road surface. The blue part in the figure is the extracted road surface.

[0055] Embodiment 3

[0056] On the basis of the process of Embodiment 1, further in step b, the following method is used to extract the road edge:

[0057] For the already extracted complete and accurate road surface data, calculate the principal component direction of the point cloud on the XY plane through principal component analysis (PCA). The goal of PCA is to find the direction of the maximum variance of the data:

[0058]

[0059] where p i is the i-th point, μ is the centroid of the point cloud, C is the covariance matrix, (p i -μ) T represents the transpose of the (p i -μ) matrix, and N_all is the number of all points. And the point cloud is aligned by rotation. After the point cloud is aligned, the road edge is extracted by slicing. During the process of aligning the point cloud, the point cloud needs to be rotated to be parallel to the slicing direction to ensure the accuracy of edge extraction;

[0060] Based on the extraction of the road edge, the road center point is extracted by calculating the geographical coordinate center of each point on the road edge in the following way.

[0061] For the effects of extracting the road edge and the center point in the embodiment, please refer to the appendix Figure 3 . Through this step, using principal component analysis can meet the rapid calculation of elements such as the edge points and center points of a long-distance complete highway. The left figure shows the extracted edge points, and the right figure shows the center points obtained from the edge points.

[0062] Embodiment 4

[0063] On the basis of the process of Embodiment 1, further in step b, the road edge is extracted in the following way:

[0064] The L1-centraline algorithm first initializes the skeleton points, and randomly selects several seed points from the point cloud data as the initial skeleton points. For each point in the point cloud, calculate its L1 distance to the skeleton points and update it according to the nearest skeleton point. The formula for the L1 distance (Manhattan distance) is:

[0065]

[0066] where p represents a point in the point cloud, and its coordinates are a three-dimensional vector (p1, p2, p3), that is, here i = 1, 2, 3, and it can also be understood as p = (x, y, z). is the j-th skeleton point, and its coordinates are three-dimensional components that is, here i = 1, 2, 3, and it can also be understood as s j = (x, y, z). d L1 (p, s j ) is the distance from point p to the skeleton point sj The L1 distance

[0067] To remove redundant skeleton points, in the embodiment, it is required that the L1 distance between each pair of skeleton points is greater than a preset threshold of 1 m. The minimum spanning tree (MST) algorithm is used to connect the thinned skeleton points. The MST connects by minimizing the total connection distance between the skeleton points:

[0068]

[0069] where E is the set of edges connecting the skeleton points, and d Euclidean (s i , s j ) is the Euclidean distance between the skeleton points s i and s j , and represents that the total connection distance e ji between the skeleton points is the shortest. The MST algorithm ensures that the total connection distance of all skeleton points is the shortest.

[0070] The L1-centraline algorithm is used to extract the centerline backbone and the B-spline curve is used to fit the point cloud distribution of the road centerline with uneven discrete distances,

[0071]

[0072] where C(u) is the curve, N i,p (u) is the basis function, P i is the control point, p is the order of the curve, n is the number of control points, and u ∈ [0, 1] is the parameter used to control the position and shape of the curve. The complete centerline with an interval of one meter in Euclidean distance is output. During the process of obtaining the road centerline, the edge line needs to be matched and a uniform distribution of point clouds with an interval of one meter is achieved to improve the accuracy and practicality of the centerline.

[0073] For the effect of the embodiment in extracting the road edge, see Appendix Figure 4 and Appendix Figure 5 , Figure 4 which is the road result of fitting the entire Pingyang section. Figure 5 It is a comparison schematic diagram of the centerline obtained by the L1-centraline algorithm and the L1 algorithm, where the red is the result obtained by the L1-centraline algorithm and the yellow is the result obtained by the L1 algorithm.

[0074] Embodiment 5

[0075] On the basis of the process of Embodiment 1, in step d, the following method is further used to achieve positioning:

[0076] For the positioning principle of the embodiment based on point cloud data to be converted to highway mileage, see AppendixFigure 6 , first, select a point that exactly matches the kilometer post. Specifically, the line connecting this point and the post should intersect the tangent of the road completely. Then, increment the post number point by point as the road advances. In this way, specific post number values are assigned to all points on the center line of the road. To avoid the accumulation of errors, after every 1000 points are located, that is, when the route advances 1 km, it is necessary to determine whether there is a large deviation between the point cloud and the next integer kilometer post. If this occurs, one of the points needs to be selected for exact matching, and then the location is cycled again to ensure the accuracy and reliability of the location.

[0077] In summary, for the highway post number location based on point cloud data element extraction proposed by the present invention, the preferred implementation process is as follows: First, through the slope method ground filtering and statistical outlier removal filtering method, extract the complete road surface, and merge the up and down roads and then separate and remove the influence of each other to obtain accurate road surface data; then, after aligning the point cloud of the already extracted complete and accurate road surface data, slice to extract the road edge. During the point cloud alignment process, the point cloud needs to be rotated to be parallel to the slice direction to ensure the accuracy of edge extraction; then, use the geographical coordinate center of each point on the road edge to preliminarily obtain the road center point, and fit to generate the continuous center line of the road. A complete center line with an interval of one meter in Euclidean distance can be output. During the process of obtaining the road center line, it is necessary to match the edge line and realize a uniform distribution of point clouds at an interval of one meter to improve the accuracy and practicality of the center line; finally, first select a point that exactly matches the kilometer post. Specifically, the line connecting this point and the post should intersect the tangent of the road completely. Then, increment the post number point by point as the road advances. To avoid the accumulation of errors, after every 1000 points are located, that is, when the route advances 1 km, it is necessary to determine whether there is a large deviation between the point cloud and the next integer kilometer post. If this occurs, one of the points needs to be selected for exact matching, and then the location is cycled again to ensure the accuracy and reliability of the location.

[0078] During specific implementation, a high-performance computing platform can be used to perform parallel processing on the point cloud data to improve the efficiency and real-time performance of data processing.

[0079] The method proposed by the present invention can achieve an accuracy of more than 1 m when positioning on roads over 100 km. The conversion results of the post number and longitude and latitude obtained in the embodiment can be seen in the following table.

[0080]

[0081] In specific implementation, the method proposed by the technical solution of the present invention can be automatically run by those skilled in the art using computer software technology. The system device for implementing the method, such as a computer-readable storage medium storing the corresponding computer program of the technical solution of the present invention and a computer device including running the corresponding computer program, should also be within the protection scope of the present invention.

[0082] Next, the highway mileage positioning device based on point cloud data element extraction provided by the present invention will be described. The highway mileage positioning device based on point cloud data element extraction described below can be correspondingly referred to the highway mileage positioning method based on point cloud data element extraction described above.

[0083] In another embodiment, the present invention provides an electronic device, which may include: a processor, a communication interface, a memory, and a communication bus. Among them, the processor, the communication interface, and the memory complete communication with each other through the communication bus. The processor can call the logical instructions in the memory to execute the highway mileage positioning method based on point cloud data element extraction.

[0084] In addition, when the logical instructions in the above-mentioned memory are implemented in the form of software functional units and sold or used as an independent product, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disc that can store program codes.

[0085] In another embodiment, the present invention further provides a computer program product. The computer program product includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the highway mileage positioning method based on point cloud data element extraction provided by the above-mentioned methods.

[0086] In another embodiment, the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it is configured to implement the highway mileage positioning method based on point cloud data element extraction provided by the above-mentioned various methods.

[0087] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative effort.

[0088] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0089] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A highway stake number positioning method based on point cloud data element extraction, characterized in that, Perform the following processing: Preprocess the original point cloud data using the slope method ground filtering and statistical outlier removal filtering methods to extract the complete road surface; Based on the extracted complete road surface data, perform point cloud alignment and then slice to extract the road edge, and preliminarily obtain the road center point using the geographical coordinate center of each point on the extracted road edge; Randomly initialize the skeleton points, iteratively optimize the positions of the skeleton points, then sparsify the skeleton points, and connect them through the minimum spanning tree to achieve the fitting generation of the continuous center line of the road; Perform the matching and positioning of the kilometer post points to achieve the accurate assignment of the mileage number.

2. The method for positioning highway mileage based on point cloud data element extraction according to claim 1, wherein: When extracting the complete road surface, for the slope method ground filtering, the ground points are selected by calculating the angle between the normal vector of each point and the vertical direction; for the statistical outlier removal filtering method, the abnormal points in the point cloud are removed by setting the threshold of the number of neighborhood points, so as to extract the complete road surface.

3. The method for positioning highway mileage based on point cloud data element extraction according to claim 1, wherein: When performing point cloud alignment, make the road direction parallel to the slice direction through rotation transformation.

4. The method for positioning highway mileage based on point cloud data element extraction according to claim 1, characterized in that: The implementation method of the fitting generation of the continuous center line of the road includes randomly selecting initial skeleton points from the road point cloud data, iteratively updating the position of each skeleton point, calculating the belonging of each point to the skeleton point using the L1 distance, and updating the skeleton point with the neighborhood median; The sparsification process removes redundant skeleton points that are too close, and connects the skeleton points using the minimum spanning tree algorithm; the B-spline curve is used to fit the complete center line.

5. The method for highway mileage positioning based on point cloud data element extraction according to claim 4, wherein: When using the B-spline curve to fit the road center line, a multi-segment fitting strategy is adopted to adapt to the complex changes in the road shape, and the fitting accuracy is optimized by the least squares method to ensure the smoothness and continuity of the center line.

6. The method for positioning highway mileage based on point cloud data element extraction according to claim 1, characterized in that: When performing the matching and positioning of the kilometer post points, an error correction mechanism is adopted. After positioning every 1000 points, the error is corrected by matching the actual kilometer post position to ensure the accuracy and reliability of the mileage number positioning.

7. The method for positioning highway mileage based on point cloud data element extraction according to claim 6, characterized in that: Input the longitude and latitude information of the integer kilometer posts of a section of highway, and obtain all the specific positions of this section with a meter-level accuracy.

8. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that: When the processor executes the program, it implements the highway mileage number positioning method based on point cloud data element extraction as described in any one of claims 1 to 7.

9. A non-transitory computer-readable storage medium storing a computer program thereon, characterized in that: When the computer program is executed by the processor, it implements the highway mileage number positioning method based on point cloud data element extraction as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that: When the computer program is executed by the processor, it implements the highway mileage number positioning method based on point cloud data element extraction as described in any one of claims 1 to 7.

Images