Rigid contact network digital model reconstruction method based on point cloud feature extraction

Abstract

The invention discloses a rigid contact network digital model reconstruction method based on point cloud feature extraction, and relates to the technical field of contact network digital model reconstruction, and the method comprises the steps: S1, collecting contact network point cloud image data in real time, and screening out a busbar-conductor point cloud profile region in the contact network point cloud image data; s2, extracting point cloud data features in the busbar-conductor point cloud profile area based on machine vision, wherein the point cloud data features comprise busbar segments, conductor arcs, conductor abrasion chord lengths, pantograph-catenary contact point conductor heights and pantograph-catenary contact point pull-out values; and S3, contact network positioning information is obtained, reconstruction coding is performed on the point cloud data features and the contact network positioning information, and a rigid contact network digital model is obtained. The extracted point cloud data features can well monitor the pantograph-catenary contact state; coded information is convenient to store and transmit, and standard data support can be provided for a subsequent data analysis system, a digital twin system and the like, so that the pantograph-catenary contact state is convenient to monitor.

Description

Technical Field

[0001] The present invention relates to the technical fields of rigid contact point cloud acquisition, rigid contact network point cloud feature extraction, and contact network digital model reconstruction, and more specifically to a rigid contact network digital model reconstruction method based on point cloud feature extraction. Background Art

[0002] With the rapid development of urban rail transit, the catenary, a core component of train power supply, has a direct impact on its operating status, impacting train safety and stability. Currently, catenary systems primarily utilize rigid or flexible contact networks. Rigid contact networks are widely used in tunnels due to their structural stability and ease of maintenance. However, variations in key rigid contact parameters such as height, pullout, and wear can lead to poor pantograph-catenary contact, potentially causing power outages, abnormal pantograph wear, and even accidents. Therefore, high-precision and efficient monitoring of rigid contact networks is crucial.

[0003] Traditional catenary monitoring relies primarily on manual measurement, such as using laser rangefinders and mechanical measuring tools. However, this method is inefficient, labor-intensive, and subject to subjective influences from the surveyor, making it difficult to meet the needs of large-scale, high-frequency inspections. In recent years, vision-based measurement methods (such as binocular vision and structured light 3D scanning) have been gradually applied to catenary monitoring. However, due to the complex lighting conditions in tunnel environments and the delicate structure of the catenary, visual methods are susceptible to interference, and measurement accuracy and stability still need to be improved.

[0004] Catenary digital model reconstruction involves the use of modern digital technologies (such as laser scanning, photogrammetry, BIM, and AI algorithms) to collect, process, and analyze high-precision three-dimensional data from the catenary system of existing electrified railways or urban rail transit. This involves collecting and monitoring catenary parameters and ultimately constructing or updating a digital model. The reconstructed catenary digital model provides data support that can be rapidly analyzed for subsequent tasks.

[0005] Existing digital model reconstructions for catenary systems are mostly targeted at flexible catenary systems. However, the structural characteristics of rigid catenary systems (such as rigid busbars and the distribution of suspension points) make feature extraction from their point cloud data more challenging. Therefore, a method for reconstructing a rigid catenary digital model is urgently needed. Summary of the Invention

[0006] To overcome the deficiencies in the aforementioned prior art, the present invention discloses a method for reconstructing a digital model of a rigid catenary network based on point cloud feature extraction. This method uses on-board equipment to acquire high-density point cloud data of the catenary network. Combined with algorithms such as noise suppression, missing segment supplementation, and circle and line fitting, it extracts point cloud data features and reconstructs and encodes the point cloud data features and catenary positioning information to produce a digital model of the rigid catenary network. This method improves the degree of measurement automation and accuracy, and the rigid catenary network digital model provides reliable data support for intelligent operation and maintenance. This method overcomes the shortcomings of the prior art in rigid catenary point cloud denoising, feature recognition, and model optimization.

[0007] In order to achieve the above objectives, the present invention adopts the following technical solutions:

[0008] A method for reconstructing a rigid contact network digital model based on point cloud feature extraction comprises the following steps:

[0009] 1. Data Collection

[0010] S1. collecting contact network point cloud image data in real time, and screening out the busbar-conductor point cloud outline area in the contact network point cloud image data;

[0011] Preferably, step S1 includes: using a vehicle-mounted device to collect contact network point cloud image data in real time, wherein the vehicle-mounted device is mounted on the roof of the vehicle, and when the vehicle is moving, the vehicle-mounted device performs point cloud imaging on the contact network according to the structured light imaging principle to obtain the contact network point cloud image data;

[0012] In the contact network point cloud image data, the horizontal coordinate w represents the pull-out value direction of the bow-catenary contact, and the vertical coordinate h represents the conductor height direction of the bow-catenary contact; in the contact network point cloud image data, the busbar-conductor point cloud contour area is screened out.

[0013] This invention utilizes a vehicle-mounted point cloud acquisition method based on the principle of structured light imaging. This method boasts high efficiency and enables continuous measurement using a high-frame-rate camera, resulting in more timely and accurate contact network data. Furthermore, this technology can capture the cross-sectional profile of the contact network, facilitating wear measurement.

[0014] 2. Feature Extraction

[0015] S2. Extracting point cloud data features in the busbar-conductor point cloud outline area based on machine vision, wherein the point cloud data features include: busbar line segments, conductor arcs, conductor wear chord lengths, pantograph-catenary contact point heights, and pantograph-catenary contact point pullout values;

[0016] Preferably, in step S2, extracting point cloud data features in the busbar-conductor point cloud outline area includes the following steps:

[0017] 2.1 Line extraction method based on noise suppression

[0018] S21. Extracting busbar segments in the busbar-conductor point cloud outline region using a line extraction method based on noise suppression;

[0019] In this invention, noise suppression is used to extract the point cloud belonging to the contact network from the original point cloud data, suppressing noise interference, which can greatly improve the accuracy of subsequent geometric parameter measurements such as wear. At the same time, suppressing noise during the feature extraction stage can reduce the time consumption of subsequent steps.

[0020] Preferably, step S21 includes the following steps:

[0021] S211, according to the prior knowledge, set the slope range (-k1, k1) and length range (d min ,d max );

[0022] S212, randomly select a point as the starting point p1 in the straight line area to be extracted in the busbar-conductor point cloud outline area, search for another point p2, and determine the equation y of the straight line formed by points p1 and p2, where the distance d between points p1 and p2 is (d min ,d max ), the slope k of the equation y of the line formed by points p1 and p2 is between (-k1, k1);

[0023] Preferably, in step S212, the distance d, slope k, and straight line equation y are respectively:

[0024]

[0025] y=k*x+b

[0026] Among them, d is the distance between point p1 and point p2, x p1 、x p2 are the horizontal coordinates of point p1 and point p2, y p1 、y p2 are the ordinates of point p1 and point p2 respectively, d min d max are the minimum and maximum distances respectively, k is the slope of the straight line equation y formed by points p1 and p2, k1 is the slope threshold, y and x are the ordinate and abscissa of the straight line equation respectively, and b is the intercept of the straight line equation.

[0027] S213. Traverse the remaining points in the busbar-conductor point cloud outline area, extract points whose distance to the line equation y is less than d1 to form a point set S1, and extract points whose distance to the line equation y is less than d2 to form a point set S2, where d1 and d2 are the first distance threshold and the second distance threshold, respectively.

[0028] S214, setting noise suppression conditions, the noise suppression conditions include: the number of points in point set S1 is greater than the number threshold n, and the ratio of the number of elements in the intersection of point set S1 and point set S2 to the number of elements in point set S2 is greater than the ratio threshold p min ;

[0029] If point set S1 and point set S2 meet the noise suppression conditions, then the points in point set S1 are represented as a straight line, the slope k′ and intercept b′ of the straight line are fitted, and the fitted straight line is used as the bus segment. The starting and ending coordinates of the bus segment are calculated.

[0030] If point set S1 and point set S2 do not meet the noise suppression conditions, this part of the data is regarded as invalid noise data and no subsequent calculation is performed; this reduces the time consumption of subsequent steps and improves the accuracy of subsequent measurements.

[0031] In the present invention, the purpose of extracting point sets S1 and S2 and setting noise suppression conditions is: for straight line segments, the collected data fluctuates and cannot be strictly judged according to the straight line conditions. By setting two different distance limits d1 and d2, the ratio of the number of points in S1 and S2 can be used to quickly determine whether the points are concentrated in the range of d1 distance, and then quickly judge the validity of the point cloud.

[0032] Preferably, in step S214, the point set S1, point set S2, noise suppression condition, slope k′ and intercept b′ are respectively:

[0033]

[0034] Among them, S1 is the point set composed of points whose distance to the straight line equation y is less than d1, S2 is the point set composed of points whose distance to the straight line equation y is less than d2, and x p 、y p are the horizontal and vertical coordinates of the remaining points traversed, k is the slope of the straight line equation y, b is the intercept of the straight line equation y, d1 is the first distance threshold, d2 is the second distance threshold, p is the ratio of the number of elements in the intersection of point set S1 and point set S2 to the number of elements in point set S2, n is the quantity threshold, p min is the ratio threshold, k′ and b′ are the slope and intercept of the straight line formed by the points in the point set S1 after satisfying the noise suppression condition, x i and y iare the horizontal and vertical coordinates of the points in the point set S1 after the noise suppression conditions are met.

[0035] S215. Repeat the above steps S211-S214 to obtain the remaining bus line segments in the bus-wire point cloud contour area, as well as the starting point and end point coordinates of the remaining bus line segments. All bus line segments are combined to form a bus line segment.

[0036] 2.2 Missing line segment supplementation

[0037] S22, supplementing the missing segments in the extracted bus segments;

[0038] Preferably, in step S22, the supplementing of the missing line segments in the extracted bus line segments includes: for the missing line segments in the bus line segments, supplementing the missing bus line segments through the endpoints of its two adjacent bus line segments.

[0039] 2.3 Obtaining the arc features of the wire

[0040] S23, using the least squares method, fitting the wire arc points in the busbar-wire point cloud outline area into a standard circle of the wire arc, and obtaining the wire arc using the standard circle of the wire arc;

[0041] Preferably, in step S23, fitting the wire arc points into a standard circle of the wire arc includes: fitting the wire arc point set S3 in the bus-wire point cloud contour area into a standard circle of the wire arc using the least squares method to obtain the center and radius of the standard circle.

[0042] Preferably, the center and radius of the standard circle are:

[0043]

[0044] Among them, x c ′、y c ' is the horizontal and vertical coordinates of the center of the standard circle, r' is the radius of the standard circle, f is the least square method, n is the number of points in the wire arc point concentration, x i 、y i are the horizontal and vertical coordinates of the points in the conductor arc point set, respectively, and S3 is the conductor arc point set in the busbar-conductor point cloud outline area.

[0045] 2.4 Obtaining the lead height, pull-out value and wear chord length

[0046] S24. Fit the conductor straight points in the busbar-conductor point cloud contour area into a straight line, and find the two intersection points of the straight line and the standard circle. The straight line between the two intersection points constitutes the conductor wear chord length, and the center point of the conductor wear chord length is used as the bow-net contact point. Calculate the bow-net contact point conductor height and bow-net contact point pull-out value.

[0047] In the present invention, steps S23 and S24 adopt a contact point and wear calculation method based on standard circle fitting, which can quickly locate the geometric center of the circle and the on-site position of the wear. The fitting method can avoid the interference of individual outliers.

[0048] 3. Feature Encoding

[0049] S3. Obtain the contact network positioning information, and reconstruct and encode the point cloud data features and the contact network positioning information to obtain a rigid contact network digital model.

[0050] In the present invention, the digital model of the road contact network model is used to reconstruct the coding. After using the coding, standard data can be formed, which is convenient for storage, transmission and query, and can provide data support that can be quickly parsed for subsequent tasks.

[0051] The reconstruction coding includes: reconstructing and coding the busbar line segment, conductor arc, conductor wear chord length, bow-net contact point height, bow-net contact point pull-out value, and contact network positioning information to obtain a rigid contact network digital model.

[0052] In the present invention, point cloud data features such as busbar segments, conductor arcs, conductor wear chord lengths, bow-net contact point heights, and bow-net contact point pull-out values ​​are extracted, and the point cloud data features and contact network positioning information are reconstructed and encoded. The extracted point cloud data features can well monitor the bow-net contact status; the encoded information is relatively convenient to store and transmit, and can provide standard data support for subsequent data analysis systems, digital twin systems, etc., thereby monitoring the bow-net contact status.

[0053] For example, the subsequent system can monitor the busbar profile by monitoring the busbar line segments; and can monitor the conductor profile by monitoring the conductor arcs.

[0054] By monitoring the wear chord length of the conductor and based on the vehicle positioning information, the degree of wear of the contact network at different points on the line can be fed back in real time. When an abnormality is detected, a real-time alarm will be issued to remind relevant departments to carry out timely maintenance to reduce unnecessary wear of the line. At the same time, by collecting a large amount of line wear data and analyzing the line wear characteristics, the line can be optimized.

[0055] By monitoring the lead height and pull-out value of the bow-catenary contact point, that is, monitoring the horizontal and vertical coordinates of the bow-catenary contact point, the points on the line can be continuously measured. When the geometric parameters are detected to deviate from the design values, an alarm can be issued for confirmation and maintenance to reduce the probability of safety accidents.

[0056] By monitoring the contact network positioning information, we can know which position of the pantograph-catenary contact status is being monitored and take corresponding measures.

[0057] Preferably, in step S3, the reconstructing and encoding the point cloud data features and the contact network positioning information includes:

[0058] [[0,h,w,c],[1,x1,y1,x2,y2],...,[2,x c ,y c ,r],[3,l,s,p,k]]

[0059] Among them, 0 represents the pantograph-catenary contact information, h, w, and c represent the pantograph-catenary contact point height, pantograph-catenary contact point pullout value, and conductor wear chord length, respectively; 1 represents the bus segment information, x1, y1, x2, and y2 represent the x- and y-axis coordinates of the starting and ending points of the bus segment, respectively, and ... represent the x- and y-axis coordinates of the starting and ending points of the remaining bus segments; 2 represents the conductor arc information, x c 、y c , r represent the coordinates of the center of the circle and the radius respectively; 3 represents the contact network positioning information, l, s, p, k represent the line, area station, pole number, and kilometer mark respectively.

[0060] Beneficial effects of the present invention:

[0061] The present invention provides a method for reconstructing a rigid catenary digital model based on point cloud feature extraction. This method collects point cloud image data of the catenary in real time and filters out the busbar-conductor point cloud outline region. Point cloud data features within this region, including busbar segments, conductor arcs, conductor wear chord lengths, bow-catenary contact point heights, and bow-catenary contact point pullout values, are then extracted. The point cloud data features and catenary positioning information are then reconstructed and encoded to produce a rigid catenary digital model. This rigid catenary digital model provides reliable data support for subsequent intelligent operation and maintenance. The extracted point cloud data features can effectively monitor the bow-catenary contact status, and the encoded information is easily stored and transmitted, providing standard data support for subsequent data analysis systems, digital twin systems, and other systems, facilitating the monitoring of bow-catenary contact status.

[0062] The rigid catenary digital model reconstruction method based on point cloud feature extraction provided by the present invention has the advantages of high efficiency, high precision, objectivity, good repeatability, strong adaptability, information integration, and digitization, as follows:

[0063] High efficiency: vehicle-mounted point cloud acquisition equipment, suitable for large-scale automated measurement;

[0064] High precision: Machine vision can detect very subtle wear changes with high accuracy;

[0065] Good objectivity and repeatability: not affected by emotions and can provide consistent measurement results;

[0066] Strong adaptability: can work in various environments, including places that are difficult to reach or dangerous;

[0067] Information integration: Complete the measurement of multiple parameters and the integration of multiple information at one time, improving the integration of data;

[0068] Digitization: The reconstructed line data can be easily digitally processed and analyzed. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 This is the process of the rigid contact network digital model reconstruction method based on point cloud feature extraction of the present invention;

[0070] Figure 2 It is the actual rigid contact network of the present invention;

[0071] Figure 3 This is the point cloud outline of the “busbar-conductor” of the present invention;

[0072] Figure 4 It is the point cloud feature representation of the present invention;

[0073] Figure 5 The wear string length and the pantograph contact point of the present invention are shown. DETAILED DESCRIPTION

[0074] The following will provide a clear and complete description of the concept, specific structure and technical effects of the present invention in conjunction with the embodiments and drawings, so as to fully understand the purpose, features and effects of the present invention.

[0075] Example 1

[0076] A method for reconstructing the digital model of rigid contact network based on point cloud feature extraction, such as Figure 1 As shown, the following steps are included:

[0077] S1. collecting contact network point cloud image data in real time, and screening out the busbar-conductor point cloud outline area in the contact network point cloud image data;

[0078] S2. Extracting point cloud data features in the busbar-conductor point cloud outline area based on machine vision, wherein the point cloud data features include: busbar line segments, conductor arcs, conductor wear chord lengths, pantograph-catenary contact point heights, and pantograph-catenary contact point pullout values;

[0079] S3. Obtain the contact network positioning information, and reconstruct and encode the point cloud data features and the contact network positioning information to obtain a rigid contact network digital model.

[0080] Example 2

[0081] This embodiment further elaborates on step S1 based on embodiment 1. Step S1 performs real-time point cloud data collection. Point cloud imaging of the contact network is performed based on the principle of structured light imaging. The equipment is installed on the roof of the vehicle and performs real-time point cloud collection while the vehicle is moving. Figure 2 is the actual rigid catenary model, Figure 3 The point cloud silhouette collected by the device.

[0082] like Figure 3 From the imaging effect diagram of the equipment, it can be seen that there is a lot of noise interference in the field of view. The part that really needs attention is the "bus-conductor", where w represents the pull-out direction of the bow-net contact, and h represents the guide height direction of the bow-net contact.

[0083] Example 3

[0084] This embodiment further elaborates on step S2 based on embodiment 2, and step S2 performs feature extraction of point cloud data. Figure 4 As shown, it is the feature information to be extracted, where l1 to l 10 Represents the 10 line segments on the busbar outline, c1 is the arc outline of the conductor, l 11 In the present invention, in order to obtain real and accurate contact network data, the locator is a part of the contact and can be measured by point cloud data, so it is fed back and can be used as analysis data for downstream tasks.

[0085] The step of extracting point cloud data features from the busbar-conductor point cloud outline area comprises the following steps:

[0086] 2.1 Line extraction method based on noise suppression

[0087] S21. Extracting busbar segments in the busbar-conductor point cloud outline region using a line extraction method based on noise suppression, comprising the following steps:

[0088] S211, according to the prior knowledge, set the slope range (-k1, k1) and length range (d min ,d max );

[0089] S212, randomly select a point as the starting point p1 in the straight line area to be extracted in the busbar-conductor point cloud outline area, search for another point p2, and determine the equation y of the straight line formed by points p1 and p2, where the distance d between points p1 and p2 is (d min ,d max ), the slope k of the equation y of the line formed by points p1 and p2 is between (-k1, k1);

[0090] The distance d, slope k, and straight line equation y are respectively:

[0091]

[0092] y=k*x+b

[0093] Among them, d is the distance between point p1 and point p2, x p1 、x p2 are the horizontal coordinates of point p1 and point p2, y p1 、y p2 are the ordinates of point p1 and point p2 respectively, d min d max are the minimum and maximum distances respectively, k is the slope of the straight line equation y formed by points p1 and p2, k1 is the slope threshold, y and x are the ordinate and abscissa of the straight line equation respectively, and b is the intercept of the straight line equation.

[0094] S213. Traverse the remaining points in the busbar-conductor point cloud outline area, extract points whose distance to the line equation y is less than d1 to form a point set S1, and extract points whose distance to the line equation y is less than d2 to form a point set S2, where d1 and d2 are the first distance threshold and the second distance threshold, respectively.

[0095] S214, setting noise suppression conditions, the noise suppression conditions include: the number of points in point set S1 is greater than the number threshold n, and the ratio of the number of elements in the intersection of point set S1 and point set S2 to the number of elements in point set S2 is greater than the ratio threshold p min ;

[0096] If point set S1 and point set S2 meet the noise suppression conditions, the points in point set S1 are represented as a straight line, the slope k′ and intercept b′ of the straight line are fitted, and the fitted straight line is used as the bus segment, and the start and end coordinates of the bus segment are calculated.

[0097] If the point set S1 and the point set S2 do not meet the noise suppression conditions, this part of the data is regarded as invalid noise data and no subsequent calculation is performed.

[0098] The point set S1, point set S2, noise suppression condition, slope k' and intercept b' are respectively:

[0099]

[0100]

[0101] Among them, S1 is the point set composed of points whose distance to the straight line equation y is less than d1, S2 is the point set composed of points whose distance to the straight line equation y is less than d2, and x p 、y pare the horizontal and vertical coordinates of the remaining points traversed, k is the slope of the straight line equation y, b is the intercept of the straight line equation y, d1 is the first distance threshold, d2 is the second distance threshold, p is the ratio of the number of elements in the intersection of point set S1 and point set S2 to the number of elements in point set S2, n is the quantity threshold, p min is the ratio threshold, k′ and b′ are the slope and intercept of the straight line formed by the points in the point set S1 after satisfying the noise suppression condition, x i and y i are the horizontal and vertical coordinates of the points in the point set S1 after the noise suppression conditions are met.

[0102] S215. Repeat the above steps S211-S214 to obtain the remaining bus line segments in the bus-wire point cloud contour area, as well as the starting point and end point coordinates of the remaining bus line segments. All bus line segments are combined to form a bus line segment.

[0103] 2.2 Missing line segment supplementation

[0104] S22. Supplementing the missing line segments in the extracted bus line segments, including: supplementing the missing line segments in the bus line segments through the endpoints of two adjacent bus line segments.

[0105] like Figure 4 As shown in the figure, due to reasons such as viewing angle, the line segment l7 in the figure is missing, and the l7 feature is supplemented by the endpoints of the buses l6 and l8.

[0106] 2.3 Obtaining the arc features of the wire

[0107] S23. Use the least squares method to fit the wire arc points in the bus-wire point cloud contour area into a standard circle of the wire arc, and use the wire arc standard circle to obtain the wire arc, including: using the least squares method to fit the wire arc point set S3 in the bus-wire point cloud contour area into a standard circle of the wire arc, and obtain the center and radius of the standard circle.

[0108] The center and radius of the standard circle are:

[0109]

[0110] Among them, x c ′、y c ' is the horizontal and vertical coordinates of the center of the standard circle, r' is the radius of the standard circle, f is the least square method, n is the number of points in the wire arc point concentration, x i 、y i are the horizontal and vertical coordinates of the points in the conductor arc point set, respectively, and S3 is the conductor arc point set in the busbar-conductor point cloud outline area.

[0111] like Figure 4As shown, the point set S3 between l5 and l6 is fitted into a circle using the least squares method to obtain the center and radius of c1. The center of the fitted circle (x c ˊ,y c ˊ) and radius rˊ.

[0112] 2.4 Obtaining the lead height, pull-out value and wear chord length

[0113] S24. Fit the conductor straight points in the busbar-conductor point cloud contour area into a straight line, and find the two intersection points of the straight line and the standard circle. The straight line between the two intersection points constitutes the conductor wear chord length, and the center point of the conductor wear chord length is used as the bow-net contact point. Calculate the bow-net contact point conductor height and bow-net contact point pull-out value.

[0114] like Figure 5 As shown, the center and radius of the standard circle are obtained through step S23. Through the positional relationship between the actual point cloud and the standard circle, the worn chord length can be calculated, and the center point of the chord length is used as the bow-net contact point, and the lead height and pull-out value are calculated.

[0115] Example 4

[0116] This embodiment further elaborates on step S3, based on embodiment 3. Step S3 reconstructs and encodes the digital model of the line contact network model. After step S2, the components of the "busbar-conductor" line segments and arcs, as well as the height of the bow-catenary contact point, the pull-out value, and the wear chord length are obtained. This information is only a cross-sectional feature. In addition to the positioning information, the bow-catenary contact situation of the entire line can be encoded. The encoding is as follows:

[0117] [[0,h,w,c],[1,x1,y1,x2,y2],...,[2,x c ,y c ,r],[3,l,s,p,k]]

[0118] Among them, the first bit is the information flag bit, 0 represents the pantograph-net contact information, h, w, c represent the pantograph-net contact point conductor height, pantograph-net contact point pullout value, and conductor wear chord length respectively; 1 represents the bus segment information, x1, y1, x2, y2 represent the x-axis and y-axis coordinates of the starting point and end point of the bus segment respectively, ... represent the x-axis and y-axis coordinates of the starting point and end point of the other bus segments; 2 represents the conductor arc information, x c 、y c , r represent the coordinates of the center of the circle and the radius respectively; 3 represents the contact network positioning information, l, s, p, k represent the line, area station, pole number, and kilometer mark respectively.

[0119] The storage and transmission of coded information is relatively convenient, and can provide standard data support for subsequent data analysis systems and digital twin systems.

[0120] The above is a detailed description of the embodiments of the present invention, but the present invention is not limited to the embodiments. Those skilled in the art may make various equivalent modifications or substitutions without departing from the spirit of the present invention. These equivalents or substitutions are all included in the scope defined by the claims of the present invention.

Claims

1. A method for reconstructing a rigid contact network digital model based on point cloud feature extraction, characterized in that: The following steps are involved: S1. collecting contact network point cloud image data in real time, and screening out the busbar-conductor point cloud outline area in the contact network point cloud image data; S2. Extracting point cloud data features in the busbar-conductor point cloud outline area based on machine vision, wherein the point cloud data features include: busbar line segments, conductor arcs, conductor wear chord lengths, pantograph-catenary contact point heights, and pantograph-catenary contact point pullout values; S3. Obtain the contact network positioning information, and reconstruct and encode the point cloud data features and the contact network positioning information to obtain a rigid contact network digital model.

2. The method for reconstructing a rigid contact network digital model based on point cloud feature extraction according to claim 1, characterized in that: Step S1 includes: using a vehicle-mounted device to collect contact network point cloud image data in real time, wherein the vehicle-mounted device is mounted on the vehicle roof, and when the vehicle is moving, the vehicle-mounted device performs point cloud imaging on the contact network according to the structured light imaging principle to obtain contact network point cloud image data; In the contact network point cloud image data, the horizontal coordinate w represents the pull-out value direction of the bow-catenary contact, and the vertical coordinate h represents the conductor height direction of the bow-catenary contact; in the contact network point cloud image data, the busbar-conductor point cloud contour area is screened out.

3. The method for reconstructing a rigid contact network digital model based on point cloud feature extraction according to claim 1, characterized in that: In step S2, extracting point cloud data features in the busbar-conductor point cloud outline area includes the following steps: S21. Extracting busbar segments in the busbar-conductor point cloud outline region using a line extraction method based on noise suppression; S22, supplementing the missing segments in the extracted bus segments; S23, using the least squares method, fitting the wire arc points in the busbar-wire point cloud outline area into a standard circle of the wire arc, and obtaining the wire arc using the standard circle of the wire arc; S24. Fit the conductor straight points in the busbar-conductor point cloud contour area into a straight line, and find the two intersection points of the straight line and the standard circle. The straight line between the two intersection points constitutes the conductor wear chord length, and the center point of the conductor wear chord length is used as the bow-net contact point. Calculate the bow-net contact point conductor height and bow-net contact point pull-out value.

4. The method for reconstructing a rigid contact network digital model based on point cloud feature extraction according to claim 3, characterized in that: Step S21 includes the following steps: S211, according to the prior knowledge, set the slope range (-k1, k1) and length range (d min ,d max ); S212, randomly select a point as the starting point p1 in the straight line area to be extracted in the busbar-conductor point cloud outline area, search for another point p2, and determine the equation y of the straight line formed by points p1 and p2, where the distance d between points p1 and p2 is (d min ,d max ), the slope k of the equation y of the line formed by points p1 and p2 is between (-k1, k1); S213. Traverse the remaining points in the busbar-conductor point cloud outline area, extract points whose distance to the line equation y is less than d1 to form a point set S1, and extract points whose distance to the line equation y is less than d2 to form a point set S2, where d1 and d2 are the first distance threshold and the second distance threshold, respectively. S214, setting noise suppression conditions, the noise suppression conditions include: the number of points in point set S1 is greater than the number threshold n, and the ratio of the number of elements in the intersection of point set S1 and point set S2 to the number of elements in point set S2 is greater than the ratio threshold p min ; If point set S1 and point set S2 meet the noise suppression conditions, then the points in point set S1 are represented as a straight line, the slope k′ and intercept b′ of the straight line are fitted, and the fitted straight line is used as the bus segment. The starting and ending coordinates of the bus segment are calculated. If the point set S1 and the point set S2 do not meet the noise suppression conditions, then this part of the data is regarded as invalid noise data and no subsequent calculation is performed; S215. Repeat the above steps S211-S214 to obtain the remaining bus line segments in the bus-wire point cloud contour area, as well as the starting point and end point coordinates of the remaining bus line segments. All bus line segments are combined to form a bus line segment.

5. The method for reconstructing a rigid contact network digital model based on point cloud feature extraction according to claim 4, characterized in that: In step S212, the distance d, slope k, and straight line equation y are respectively: Among them, d is the distance between point p1 and point p2, x p1 、x p2 are the horizontal coordinates of point p1 and point p2, y p1 、y p2 are the ordinates of point p1 and point p2 respectively, d min d max are the minimum and maximum distances respectively, k is the slope of the straight line equation y formed by points p1 and p2, k1 is the slope threshold, y and x are the ordinate and abscissa of the straight line equation respectively, and b is the intercept of the straight line equation.

6. The method for reconstructing a rigid contact network digital model based on point cloud feature extraction according to claim 4, characterized in that: In step S214, the point set S1, point set S2, noise suppression condition, slope k' and intercept b' are respectively: Among them, S1 is the point set composed of points whose distance to the straight line equation y is less than d1, S2 is the point set composed of points whose distance to the straight line equation y is less than d2, and x p 、y p are the horizontal and vertical coordinates of the remaining points traversed, k is the slope of the straight line equation y, b is the intercept of the straight line equation y, d1 is the first distance threshold, d2 is the second distance threshold, p is the ratio of the number of elements in the intersection of point set S1 and point set S2 to the number of elements in point set S2, n is the quantity threshold, p min is the ratio threshold, k′ and b′ are the slope and intercept of the straight line formed by the points in the point set S1 after satisfying the noise suppression condition, x i and y i are the horizontal and vertical coordinates of the points in the point set S1 after the noise suppression conditions are met.

7. The method for reconstructing a rigid contact network digital model based on point cloud feature extraction according to claim 3, characterized in that: In step S22, the supplementing of the missing line segments in the extracted bus line segments includes: for the missing line segments in the bus line segments, supplementing the missing bus line segments through the endpoints of two adjacent bus line segments.

8. The method for reconstructing a rigid contact network digital model based on point cloud feature extraction according to claim 3, characterized in that: In step S23, fitting the wire arc points into a standard circle of the wire arc includes: fitting the wire arc point set S3 in the busbar-wire point cloud contour area into a standard circle of the wire arc using the least squares method to obtain the center and radius of the standard circle.

9. The method for reconstructing a rigid catenary digital model based on point cloud feature extraction according to claim 8, characterized in that: The center and radius of the standard circle are: Among them, x c ′、y c ' is the horizontal and vertical coordinates of the center of the standard circle, r' is the radius of the standard circle, f is the least square method, n is the number of points in the wire arc point concentration, x i 、y i are the horizontal and vertical coordinates of the points in the conductor arc point set, respectively, and S3 is the conductor arc point set in the busbar-conductor point cloud outline area.

10. The method for reconstructing a rigid contact network digital model based on point cloud feature extraction according to claim 1, characterized in that: In step S3, the reconstructing and encoding of the point cloud data features and the contact network positioning information includes: [[0,h,w,c],[1,x1,y1,x2,y2],...,[2,x c ,y c ,r],[3,l,s,p,k]] Among them, 0 represents the pantograph-catenary contact information, h, w, and c represent the pantograph-catenary contact point height, pantograph-catenary contact point pullout value, and conductor wear chord length, respectively; 1 represents the bus segment information, x1, y1, x2, and y2 represent the x- and y-axis coordinates of the starting and ending points of the bus segment, respectively, and ... represent the x- and y-axis coordinates of the starting and ending points of the remaining bus segments; 2 represents the conductor arc information, x c 、y c , r represent the coordinates of the center of the circle and the radius respectively; 3 represents the contact network positioning information, l, s, p, k represent the line, area station, pole number, and kilometer mark respectively.

Images