Deep learning-based railway track vertex cloud data extraction method and system

Abstract

The invention discloses a railway track vertex cloud data extraction method and system based on deep learning. The method comprises the following steps: classifying track point cloud data by using a point cloud network enhanced multi-scale classification model to obtain a track key component point cloud classification data set; rail head point cloud data are extracted from the rail key component point cloud classification data set, center line fitting is carried out, a rail head center line is generated, and a reference point is selected on the rail head center line; a rolling window method is adopted, and a plurality of slices are generated along the rail head center line according to the reference points; and carrying out rail top point extraction on the plurality of slices to obtain a continuous rail top coordinate point set. The system comprises a data classification module, a rail head center line generation module, a slice generation module and a rail vertex extraction module. The rail top data extraction efficiency can be improved, the data accuracy and the rail safety are ensured, and scientific maintenance and management of the railway rail are facilitated.

Description

Technical Field

[0001] The present invention relates to the field of railway engineering technology, and in particular to a method and system for extracting railway track vertex cloud data based on deep learning. Background Art

[0002] In recent years, with the continuous upgrading and expansion of my country's railway infrastructure, particularly the construction and operation of high-speed and heavy-haul railways, the operational quality and safety of railway lines have been subject to increasingly stringent requirements. The rail top is a key component in the track structure, where the train directly contacts the track. Its condition has a significant impact on the safety, stability, and comfort of train operation. However, due to its long service life and frequent train operation, the track is prone to varying degrees of deformation and wear, posing a higher technical challenge to railway maintenance.

[0003] Track point cloud data is primarily acquired through a variety of measurement methods, including laser scanning, satellite positioning, and inertial measurement. These technologies can accurately determine the three-dimensional coordinates of the track. However, during the track measurement process, various factors, such as unstable satellite signals, vibrations during laser measurement equipment, and the complexity of the track environment, often lead to noise and errors in some data. If this data is not effectively processed, these errors will seriously affect the accuracy of rail top extraction, further compromising subsequent track maintenance and deformation analysis.

[0004] To ensure the safety and operational quality of railway tracks, it is necessary to accurately extract the coordinates of the rail tops to analyze track deformation. Traditional rail top extraction methods rely on manual inspection and simple mathematical models. These methods are inefficient and lack accuracy when processing large-scale, highly complex track point cloud data. With the rapid development of deep learning technology, deep learning-based point cloud data processing methods are gradually being applied to track engineering, enabling more automated and precise track analysis. Summary of the Invention

[0005] In view of this, the purpose of the embodiments of the present invention is to provide a railway track top cloud data extraction method and system based on deep learning, which can improve the efficiency of rail top data extraction, ensure the accuracy of data and the safety of tracks, and contribute to the scientific maintenance and management of railway tracks.

[0006] In a first aspect, an embodiment of the present invention provides a railway track vertex cloud data extraction based on deep learning, which includes: The track point cloud data is classified using a point cloud network enhanced multi-scale classification model to obtain a point cloud classification dataset of key track components.

[0007] The rail head point cloud data is extracted from the point cloud classification data set of the key rail components, and centerline fitting is performed to generate a rail head centerline, and a reference point is selected on the rail head centerline.

[0008] A rolling window method is used to generate a number of slices along the rail head centerline according to the reference point.

[0009] Track vertex extraction is performed on a number of the slices to obtain a continuous track top coordinate point set.

[0010] In combination with the first aspect, an embodiment of the present invention provides a first possible implementation of the first aspect, wherein the step of classifying the track point cloud data using the point cloud network enhanced multi-scale classification model to obtain a track key component point cloud classification dataset includes: The discrete track point cloud data is acquired through laser scanning equipment, and the discrete track point cloud data is input into the point cloud network enhanced multi-scale classification model.

[0011] Multi-scale feature extraction is performed on the discrete track point cloud data using the point cloud network enhanced multi-scale classification model to capture the features of different key track components.

[0012] The discrete track point cloud data is divided into several categories of point cloud data, where the categories include at least one of a rail head, a rail body, a fastener, and a rail plate.

[0013] The point cloud data of different categories are aggregated respectively to obtain a point cloud classification dataset of key track components.

[0014] In combination with the first aspect, an embodiment of the present invention provides a second possible implementation of the first aspect, wherein the step of performing multi-scale feature extraction on the discrete track point cloud data using the point cloud network enhanced multi-scale classification model to capture features of different key track components includes: For each point in the point cloud , define the neighborhood point set .

[0015] Calculate each of the points and its neighboring points Geometric features, extract local features ,in, The neighborhood point To the point The feature calculation results.

[0016] In combination with the first aspect, an embodiment of the present invention provides a third possible implementation of the first aspect, wherein extracting rail head point cloud data from the rail key component point cloud classification dataset, performing centerline fitting to generate a rail head centerline, and selecting a reference point on the rail head centerline includes: The rail head point cloud data is extracted from the rail key component point cloud classification data set, where the rail head point cloud data includes the spatial coordinates (X, Y, Z) of the rail head.

[0017] A local weighted regression method is used to assign different weights to the point cloud data to capture the local features of the rail head.

[0018] The fitting points are interpolated using multiple spline interpolation method to generate a continuous rail head centerline.

[0019] On the center line of the rail head, equally spaced points are selected as reference points.

[0020] In combination with the first aspect, an embodiment of the present invention provides a fourth possible implementation of the first aspect, wherein the using of a local weighted regression method to assign different weights to point cloud data to capture local features of the rail head includes: For each of the fitting points of the rail head point cloud data, use the fitting function Calculate the weight of its neighborhood points, where is a smoothing parameter used to control the weighted range of neighborhood points.

[0021] In combination with the first aspect, an embodiment of the present invention provides a fifth possible implementation of the first aspect, wherein the interpolating the fitting points using a multiple spline interpolation method to generate a continuous rail head centerline includes: Assume that the discrete fitting point set of the rail head centerline is , each point Indicates the specific spatial coordinates on the center line of the rail head, and each point as an interpolation node.

[0022] In the interval between every two adjacent interpolation nodes Define the cubic spline function ,in, is the coefficient of the i-th segment of the spline, and each segment of the spline In the said interval Valid within.

[0023] Set the interpolation conditions, each segment of the spline at each node The value at is equal to the known coordinates of the fitting point .

[0024] Set the continuity condition of the first-order derivative at the node The slope of the spline in the first section is equal to the slope of the spline in the second section. .

[0025] Set the continuity condition of the second-order derivative, the curvature of the spline described in the first paragraph is equal to the curvature of the spline described in the second paragraph ; Set the natural boundary conditions so that the second derivative of the cubic spline function curve at both ends is zero ; The interpolation condition, the continuity condition of the first-order derivative, the continuity condition of the second-order derivative and the natural boundary condition are constructed into a linear equation, and the coefficients of each segment of the spline are obtained by solving the equation. .

[0026] Each segment of the spline Combined together, a continuous rail head center line is obtained.

[0027] In combination with the first aspect, an embodiment of the present invention provides a sixth possible implementation of the first aspect, wherein the step of generating a plurality of slices along the centerline of the rail head according to the reference point using a rolling window method includes: According to the selected reference points, several main slices are generated on the center line of the rail head.

[0028] By formula Subdivision slices are generated between adjacent main slices, wherein: is the normal vector of the subdivision slice plane, which is parallel or perpendicular to the direction vector of the rail head centerline, is the position vector of a point on the center line of the rail head, and the subdivision slice is centered at this position, is the position vector of any point within the subdivided slice.

[0029] The subdivided slices are generated sequentially between the main slices in a rolling manner along the center line of the rail head.

[0030] In combination with the first aspect, an embodiment of the present invention provides a seventh possible implementation of the first aspect, wherein extracting rail vertices from the plurality of slices to obtain a continuous set of rail top coordinate points includes: For each point cloud data in the slice, the point with the highest height is selected as the vertex of the track of the slice according to the height value of each point.

[0031] According to the order of the slices on the center line of the rail head, the coordinates of the rail vertex of each slice are arranged in the direction of the railway to form a rail top line.

[0032] The coordinates of all the extracted rail vertices are connected in sequence to generate a continuous rail top coordinate point set.

[0033] In combination with the first aspect, an embodiment of the present invention provides an eighth possible implementation of the first aspect, wherein, for each of the point cloud data in the slice, selecting the point with the highest height as the vertex of the track of the slice according to the height value of each point includes: For each of the slices, traverse all point cloud data and obtain the height value of each point .

[0034] The height value The height of the highest corresponding point is defined as the rail top height , The three-dimensional coordinates (X, Y, Z) of the point are recorded as the vertex of the track of the slice.

[0035] In a second aspect, an embodiment of the present invention further provides a railway track vertex cloud data extraction system based on deep learning, which includes: The data classification module is used to classify the track point cloud data using the point cloud network enhanced multi-scale classification model to obtain a point cloud classification dataset of key track components.

[0036] The rail head centerline generation module is used to extract the rail head point cloud data from the rail key component point cloud classification data set, perform centerline fitting, generate the rail head centerline, and select a reference point on the rail head centerline.

[0037] The slice generation module is used to generate a plurality of slices along the center line of the rail head according to the reference point by using a rolling window method.

[0038] The rail vertex extraction module is used to extract rail vertices from a plurality of the slices to obtain a continuous rail top coordinate point set.

[0039] The beneficial effects of the present invention are: This invention provides a fully automated rail head vertex extraction method that doesn't rely on human intervention, significantly reducing the need for manual intervention and the potential for human error. By automatically loading point cloud data, performing centerline fitting, and extracting vertices, the entire process is highly integrated and automated, making track inspection more efficient and faster, meeting the high-efficiency requirements of modern railway maintenance.

[0040] Using local weighted regression and multiple spline interpolation to fit the centerline of the rail head point cloud data helps ensure centerline accuracy and reduces deviations caused by point cloud noise and abnormal data. Furthermore, a fixed-interval slice analysis method accurately identifies the highest point within each slice as the rail vertex, ensuring high accuracy and reliability of the extracted points. This high-precision extraction method effectively improves the accuracy of track flatness assessment and deformation monitoring.

[0041] This method uses a method of slicing at regular intervals along the centerline to ensure that rail vertices are evenly distributed across the track. Uniform vertex extraction is crucial for evaluating track flatness and effectively avoids bias in track condition assessments caused by uneven sampling. Evenly distributed vertex data provides a scientific basis for long-term track monitoring and maintenance, more comprehensively reflecting the actual track condition and supporting accurate assessment of track deformation.

[0042] The method presented in this paper is highly adaptable, not only to measuring different track shapes and lengths, but also to addressing various point cloud data accuracy and density requirements. By flexibly adjusting the number of centerline fitting points and the intervals between slice analyses, it can adapt to track inspection tasks of varying types and conditions, demonstrating its broad potential for application.

[0043] This invention provides a scientific and effective method for the automated extraction of rail head vertices. The research results are of great significance to the inspection and maintenance of railway tracks. Through highly automated and high-precision extraction methods, the efficiency and quality of track measurement and evaluation are greatly improved, providing important technical support for the healthy and safe development of the railway industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without paying any creative work.

[0045] Figure 1 This is a flow chart of the railway track vertex cloud data extraction method based on deep learning of the present invention; Figure 2 Schematic diagram of the classification of track point cloud data in the railway track vertex cloud data extraction method based on deep learning of the present invention; Figure 3 Schematic diagram of multiple spline interpolation of the rail head centerline in the railway rail vertex cloud data extraction method based on deep learning of the present invention; Figure 4 A schematic diagram of generating slices along the centerline of the rail head in the railway rail vertex cloud data extraction method based on deep learning of the present invention; Figure 5 Schematic diagram of rail vertex extraction within a cross-section slice in the railway rail vertex cloud data extraction method based on deep learning of the present invention; Figure 6 This is a schematic diagram of the generation of the rail top coordinate point set and the rail top continuous line shape in the railway rail vertex cloud data extraction method based on deep learning of the present invention; Figure 7 This is a schematic diagram of the extraction process of the railway track vertex cloud data extraction method based on deep learning in the present invention. DETAILED DESCRIPTION

[0046] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0047] This paper proposes a method for extracting railway rail vertex cloud data based on deep learning and rolling window slicing. Using the PointNet++ MSG model, the track point cloud is classified to accurately extract rail head data. This method, combined with centerline fitting and rolling window slicing analysis techniques, ensures high-precision extraction of rail vertices. This method not only improves the efficiency of rail top data extraction, but also ensures data accuracy and track safety, contributing to the scientific maintenance and management of railway tracks.

[0048] Please refer to Figures 1 to 7 A first embodiment of the present invention provides a railway rail vertex cloud data extraction method based on deep learning, which includes: using a point cloud network enhanced multi-scale classification model to classify track point cloud data to obtain a track key component point cloud classification dataset; extracting rail head point cloud data from the track key component point cloud classification dataset, performing centerline fitting to generate a rail head centerline, and selecting a reference point on the rail head centerline; using a rolling window method to generate a plurality of slices along the rail head centerline based on the reference point; and performing rail vertex extraction on the plurality of slices to obtain a continuous rail top coordinate point set.

[0049] Among them, the use of the point cloud network enhanced multi-scale classification model to classify the track point cloud data to obtain a point cloud classification dataset of key track components includes: acquiring discrete track point cloud data through laser scanning equipment, and inputting the discrete track point cloud data into the point cloud network enhanced multi-scale classification model; performing multi-scale feature extraction on the discrete track point cloud data through the point cloud network enhanced multi-scale classification model to capture the features of different key track components; dividing the discrete track point cloud data into several categories of point cloud data, the categories including at least one of rail head, rail body, fastener and rail plate; and aggregating the point cloud data of different categories to obtain a point cloud classification dataset of key track components.

[0050] Specifically, the rail head has a flat-top geometric shape, and the point cloud network enhanced multi-scale classification model distinguishes it by identifying horizontal structural features through neighborhood point relationships; the rail body is long and vertical, adjacent to the rail head, and is distinguished by longitudinal features; the track plate is distinguished by its planarity extracted through multi-scale information; the fastener is smaller in volume than the rail head, the rail body and the track plate, and is distinguished by being close to the track plate in position but relatively independent in shape.

[0051] The multi-scale feature extraction of the discrete track point cloud data by the point cloud network enhanced multi-scale classification model to capture the features of different key track components includes: for each point in the point cloud , define the neighborhood point set ; Calculate each of the points and its neighboring points Geometric features, extract local features ,in, The neighborhood point To the point By aggregating these local features, the various components of the track can be better classified, such as Figure 2 shown.

[0052] The method further comprises extracting rail head point cloud data from the rail key component point cloud classification data set, performing centerline fitting, generating a rail head centerline, and selecting reference points on the rail head centerline, wherein the method comprises: extracting rail head point cloud data from the rail key component point cloud classification data set, the rail head point cloud data including the spatial coordinates (X, Y, Z) of the rail head; assigning different weights to the point cloud data using a local weighted regression method to capture local features of the rail head; interpolating the fitting points using a multiple spline interpolation method to generate a continuous rail head centerline; and selecting equally spaced points on the rail head centerline as reference points.

[0053] The method of using a local weighted regression method to assign different weights to the point cloud data to capture the local features of the rail head includes: for each fitting point of the rail head point cloud data, using a fitting function Calculate the weight of its neighborhood points, where is a smoothing parameter used to control the weighted range of neighborhood points, ensuring that points closer to the center point have a greater impact on the fitting results.

[0054] The method of interpolating the fitting points using multiple spline interpolation to generate a continuous rail head center line includes: assuming that the discrete fitting point set of the rail head center line is , each point Indicates the specific spatial coordinates on the center line of the rail head, and each point As interpolation nodes, they constitute the dividing points of multi-segment splines; the interval between every two adjacent interpolation nodes Define the cubic spline function ,in, is the coefficient of the i-th segment of the spline, and each segment of the spline In the said interval Valid within; set the interpolation conditions, each segment of the spline at each node The value at is equal to the known coordinates of the fitting point , ensure that the function value of each spline segment is consistent with the discrete fitting point of the track head at the interpolation node; set the continuity condition of the first-order derivative at the node The slope of the spline in the first section is equal to the slope of the spline in the second section. ; Set the continuity condition of the second-order derivative, the curvature of the spline in the first paragraph is equal to the curvature of the spline in the second paragraph ;Set natural boundary conditions, the second-order derivative of the cubic spline function curve at both ends is zero This condition makes the rate of change of the curve at the boundary smooth, and the interpolation curve obtained transitions naturally at both ends of the rail head centerline; the interpolation condition, the continuity condition of the first-order derivative, the continuity condition of the second-order derivative and the natural boundary condition are constructed into a linear equation, and the coefficient of each segment of the spline is obtained by solving it. ; Each segment of the spline Combined, a continuous rail head center line is obtained, such as Figure 3 As shown, in In the interval, the rail head centerline is composed of these piecewise splines, which meet the requirements of smoothness and continuity in each interpolation node and its interval area.

[0055] Wherein, the rolling window method is used to generate a plurality of slices along the center line of the rail head according to the reference points, including: generating a plurality of main slices on the center line of the rail head according to the selected reference points, wherein the reference points are evenly distributed and the interval distance can be adjusted according to actual needs to ensure the smooth transition and data density of the track; and Subdivision slices are generated between adjacent main slices, wherein: is the normal vector of the subdivision slice plane, which is parallel or perpendicular to the direction vector of the rail head centerline, is the position vector of a point on the center line of the rail head, and the subdivision slice is centered at this position, is the position vector of any point within the subdivision slice. The generated subdivision slice will be parallel to the local plane of the rail head at that point, ensuring that the subdivision slice can be perpendicular to the centerline of the rail head, thereby maximizing the consistency of the rail direction. The number and spacing of subdivision slices can be set according to the requirements of rail top feature extraction. Usually, dense subdivision slices are selected to increase the resolution of rail top data, such as Figure 4 As shown, by generating several slices between every two centerline points, the extraction density of rail top coordinates can be effectively increased, thereby ensuring high resolution and high continuity of rail top data; the subdivided slices are sequentially generated between the main slices in a rolling manner along the rail head centerline. The moving step size and window size of the rolling window can be adjusted according to the rail shape and measurement accuracy requirements. Each generated subdivided slice not only contains local rail head geometric information, but also forms a continuous rail top dataset with adjacent slices, providing a complete rail top surface profile.

[0056] Each generated slice serves as a separate planar region for further analysis of the rail head point cloud's geometric features. This slice generation method ensures detailed analysis of the rail head at various locations, providing a deeper understanding of the track's details. This process provides the data foundation for accurate extraction of the rail top.

[0057] The method of extracting rail vertices from the plurality of slices to obtain a continuous rail top coordinate point set includes: selecting the point with the highest height as the rail top of the slice according to the height value of each point in the point cloud data of each slice; arranging the coordinates of the rail top of each slice in the direction of the railway according to the order of the slices on the center line of the rail head to form a rail top line; and connecting the coordinates of all the extracted rail vertices in sequence to generate a continuous rail top coordinate point set. The sequential arrangement of the rail top coordinate point set ensures the continuity and accuracy of the rail top data, thereby accurately reflecting the shape of the track. Figure 6 shown.

[0058] Wherein, for each of the point cloud data in the slice, according to the height value of each point, selecting the point with the highest height as the vertex of the slice includes: for each of the slices, traversing all the point cloud data, obtaining the height value of each point ; The height value The height of the highest corresponding point is defined as the rail top height , record the three-dimensional coordinates (X, Y, Z) of the point as the vertex of the slice, such as Figure 5 As shown, by comparing the heights of all points in each slice, the point with the largest height is extracted to ensure the accuracy of the rail vertex.

[0059] A second embodiment of the present invention provides a railway rail vertex cloud data extraction system based on deep learning, which includes: a data classification module, which is used to classify rail point cloud data using a point cloud network enhanced multi-scale classification model to obtain a rail key component point cloud classification data set; a rail head centerline generation module, which is used to extract rail head point cloud data from the rail key component point cloud classification data set, perform centerline fitting, generate a rail head centerline, and select a reference point on the rail head centerline; a slice generation module, which is used to use a rolling window method to generate a plurality of slices along the rail head centerline according to the reference point; and a rail vertex extraction module, which is used to perform rail vertex extraction on a plurality of the slices to obtain a continuous rail top coordinate point set.

[0060] The embodiments of the present invention are intended to protect a method and system for extracting railway track vertex cloud data based on deep learning, which has the following effects: To address the current demand for rail head vertex extraction on railway tracks, this paper proposes an automated rail head vertex extraction method based on deep learning and rolling window slicing. First, rail head point cloud data is acquired through deep learning. Then, the center line of the rail head point cloud data is fitted using local weighted regression and multiple spline interpolation to ensure centerline accuracy. Finally, rail vertices are extracted based on regularly spaced slice analysis to ensure a uniform distribution of vertices along the track.

[0061] The computer program product of the railway track vertex cloud data extraction method and device based on deep learning provided in the embodiments of the present invention includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the method in the previous method embodiment. The specific implementation can be found in the method embodiment and will not be repeated here.

[0062] Specifically, the storage medium can be a general storage medium, such as a mobile disk, a hard disk, etc. When the computer program on the storage medium is run, it can execute the above-mentioned railway rail vertex cloud data extraction method based on deep learning, thereby improving the efficiency of rail top data extraction and ensuring the accuracy of the data and the safety of the track.

[0063] If the functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage media include various media capable of storing program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0064] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the above-described embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above-described embodiments within the technical scope disclosed by the present invention, or replace some of the technical features therein with equivalents. Such modifications, changes, or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A railway track vertex cloud data extraction method based on deep learning, characterized in that: include: The track point cloud data is classified using the point cloud network enhanced multi-scale classification model to obtain a point cloud classification dataset of key track components. Extracting rail head point cloud data from the rail key component point cloud classification data set, performing centerline fitting, generating a rail head centerline, and selecting a reference point on the rail head centerline; Using a rolling window method, a plurality of slices are generated along the center line of the rail head according to the reference point; The track vertices of the plurality of slices are extracted to obtain a continuous track top coordinate point set.

2. The railway track vertex cloud data extraction method based on deep learning according to claim 1 is characterized in that: The point cloud network enhanced multi-scale classification model is used to classify the track point cloud data to obtain a point cloud classification data set of key track components, including: Acquire discrete track point cloud data through laser scanning equipment, and input the discrete track point cloud data into the point cloud network enhanced multi-scale classification model; Performing multi-scale feature extraction on the discrete track point cloud data through the point cloud network enhanced multi-scale classification model to capture the features of different key track components; Classifying the discrete track point cloud data into several categories of point cloud data, wherein the categories include at least one of a rail head, a rail body, a fastener, and a rail plate; The point cloud data of different categories are aggregated respectively to obtain a point cloud classification data set of key track components.

3. The railway track vertex cloud data extraction method based on deep learning according to claim 2 is characterized in that: The multi-scale feature extraction of the discrete track point cloud data by the point cloud network enhanced multi-scale classification model to capture the features of different key track components includes: For each point in the point cloud , define the neighborhood point set ; Calculate each of the points and its neighboring points Geometric features, extract local features ,in, is the neighborhood point To the point The characteristic calculation results of .

4. The railway track vertex cloud data extraction method based on deep learning according to claim 1 is characterized in that: The extracting of the rail head point cloud data from the rail key component point cloud classification data set, performing centerline fitting, generating the rail head centerline, and selecting a reference point on the rail head centerline comprises: Extracting rail head point cloud data from the rail key component point cloud classification data set, wherein the rail head point cloud data includes the spatial coordinates (X, Y, Z) of the rail head; The local weighted regression method is used to assign different weights to the point cloud data to capture the local features of the rail head; The fitting points are interpolated using multiple spline interpolation method to generate a continuous rail head centerline; On the center line of the rail head, equally spaced points are selected as reference points.

5. The railway track vertex cloud data extraction method based on deep learning according to claim 4 is characterized in that: The local weighted regression method is used to assign different weights to the point cloud data, and the local features of the rail head are captured, including: For each of the fitting points of the rail head point cloud data, a fitting function is used. Calculate the weight of its neighborhood points, where is a smoothing parameter used to control the weighted range of neighborhood points.

6. The railway track vertex cloud data extraction method based on deep learning according to claim 4 is characterized in that: The method of interpolating the fitting points using multiple spline interpolation method to generate a continuous rail head centerline includes: Assume that the discrete fitting point set of the rail head centerline is , each point Represents the specific spatial coordinates on the center line of the rail head, and each point As an interpolation node; In the interval between every two adjacent interpolation nodes Define the cubic spline function ,in, is the coefficient of the i-th spline, each spline In the said interval Internally valid; Set the interpolation conditions, each segment of the spline at each node The value at is equal to the known coordinates of the fitting point ; Set the continuity condition for the first-order derivative at the node At this point, the slope of the spline in the first section is equal to the slope of the spline in the second section. ; Set the continuity condition of the second-order derivative, the curvature of the spline described in the first paragraph is equal to the curvature of the spline described in the second paragraph ; Set natural boundary conditions, the second-order derivative of the cubic spline function curve at both ends is zero ; The interpolation condition, the continuity condition of the first-order derivative, the continuity condition of the second-order derivative and the natural boundary condition are constructed into a linear equation, and the coefficients of each segment of the spline are obtained by solving the equation. ; Each segment of the spline Combined together, a continuous rail head center line is obtained.

7. The railway track vertex cloud data extraction method based on deep learning according to claim 1 is characterized in that: The method of using the rolling window method to generate a plurality of slices along the center line of the rail head according to the reference point comprises: According to the selected reference points, a plurality of main slices are generated on the center line of the rail head; By formula Subdivided slices are generated between adjacent main slices, wherein: is the normal vector of the subdivided slice plane, parallel or perpendicular to the direction vector of the centerline of the rail head, is the position vector of a point on the centerline of the rail head, and the subdivision slice is centered at that position, is the position vector of any point in the subdivided slice; The subdivided slices are generated sequentially between the main slices in a rolling manner along the center line of the rail head.

8. The railway track vertex cloud data extraction method based on deep learning according to claim 1, characterized in that: The step of extracting rail vertices from a plurality of the slices to obtain a continuous rail top coordinate point set comprises: For each of the point cloud data in the slice, according to the height value of each point, select the point with the highest height as the vertex of the track of the slice; According to the order of the slices on the center line of the rail head, the coordinates of the rail vertex of each slice are arranged in the direction of the railway to form a rail top line; The coordinates of all the extracted rail vertices are connected in sequence to generate a continuous rail top coordinate point set.

9. The railway track vertex cloud data extraction method based on deep learning according to claim 8, characterized in that: The step of selecting the point with the highest height as the vertex of the track of each slice according to the height value of each point in the point cloud data in each slice comprises: For each of the slices, traverse all point cloud data and obtain the height value of each point ; The height value The height of the highest corresponding point is defined as the rail top height , The three-dimensional coordinates (X, Y, Z) of the point are recorded as the vertex of the track of the slice.

10. A railway track vertex cloud data extraction system based on deep learning, characterized in that: include: A data classification module is used to classify the track point cloud data using a point cloud network enhanced multi-scale classification model to obtain a point cloud classification dataset of key track components; A rail head centerline generation module is used to extract rail head point cloud data from the rail key component point cloud classification data set, perform centerline fitting, generate the rail head centerline, and select a reference point on the rail head centerline; A slice generation module, used to generate a plurality of slices along the center line of the rail head according to the reference point by using a rolling window method; The rail vertex extraction module is used to extract rail vertices from a plurality of the slices to obtain a continuous rail vertex coordinate point set.

Images