CN113870123A - Automatic detection method for contact net leading height and pulling value based on vehicle-mounted mobile laser point cloud - Google Patents

Automatic detection method for contact net leading height and pulling value based on vehicle-mounted mobile laser point cloud Download PDF

Info

Publication number
CN113870123A
CN113870123A CN202110955517.6A CN202110955517A CN113870123A CN 113870123 A CN113870123 A CN 113870123A CN 202110955517 A CN202110955517 A CN 202110955517A CN 113870123 A CN113870123 A CN 113870123A
Authority
CN
China
Prior art keywords
line
point
point cloud
contact
steel rail
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202110955517.6A
Other languages
Chinese (zh)
Other versions
CN113870123B (en
Inventor
许磊
豆孝磊
张冠军
牟春霖
巩健
谢春喜
刘成
杨元维
梁永
石德斌
董延喜
赵梦杰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China Railway Design Corp
Original Assignee
China Railway Design Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by China Railway Design Corp filed Critical China Railway Design Corp
Priority to CN202110955517.6A priority Critical patent/CN113870123B/en
Priority to PCT/CN2021/124110 priority patent/WO2023019709A1/en
Publication of CN113870123A publication Critical patent/CN113870123A/en
Application granted granted Critical
Publication of CN113870123B publication Critical patent/CN113870123B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G06T5/70—
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10028—Range image; Depth image; 3D point clouds
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30241—Trajectory

Abstract

The invention discloses a method for automatically detecting a leading-up value and a pulling-out value of a contact net based on vehicle-mounted mobile laser point cloud, which comprises the steps of firstly, setting an angle segmentation threshold according to a scanning angle corresponding to a track structure and a contact net structure, segmenting to obtain the track point cloud and the contact net point cloud, and removing point cloud noise by using a statistical filtering algorithm; secondly, based on the trajectory line of the mobile scanning, a cuboid segmentation algorithm is adopted to segment the steel rail point cloud, and a cylindrical segmentation algorithm is adopted to segment the contact line and the locating point cloud; thirdly, registering the steel rail point cloud by using PCA and GICP, extracting a contact line point cloud by using RANSAC, and extracting a suspension point cloud by using a dimensional analysis method; and finally, calculating the lead height and the pull-out value of the contact line according to the lead height and the pull-out value definitions. The method has the characteristics of high precision, high efficiency, high automation degree and the like, can be used as effective supplement of a contact net lead-up and pull-out value normalized detection mode, and serves for newly-built railway static acceptance and existing railway line operation detection.

Description

Automatic detection method for contact net leading height and pulling value based on vehicle-mounted mobile laser point cloud
Technical Field
The invention relates to the field of rapid comprehensive detection of rail transit, in particular to a method for automatically detecting the leading height and the pull-out value of a contact network based on vehicle-mounted mobile laser point cloud.
Background
The railway is an important infrastructure of the country, and the aorta of national economy is in the backbone position in a comprehensive transportation system. In 1975, the electrified transformation of the whole line of the Baozhu line is completed, and the Baozhu line becomes the first electrified railway in China. After that, the electrified railway in China is developed rapidly, and the total mileage of railway operation in China exceeds 14.63 kilometers by 2020 years, wherein the high-speed railway reaches 3.8 kilometers, the electrified mileage is 10.6 kilometers, and the electrification rate is 72.8%. The contact net is used as the core part of the electrified railway, and plays an important role in the aspects of ensuring the railway transportation safety, improving the transportation efficiency, reducing the transportation energy consumption and the like.
The contact net is placed in an open air environment, and is easy to age and change geometric shape and position in a severe environment. During train operation, due to the reasons of irregular operation, abnormal relation between bow nets, electric corrosion and the like, the contact net can deform or even break. When the contact network breaks down, the operation of the whole railway line is influenced because no standby line exists, thereby causing huge economic loss and adverse social influence. The geometric parameters of the contact network mainly comprise contact line height guide and pull-out values, wherein the height guide refers to the vertical distance from the bottom surface of the contact line to the rail surface, and the pull-out value refers to the offset of the contact line to the center line of the pantograph at a positioning point. For a newly-built contact network project, the lead-height and pull-out value overrun caused by installation errors is timely discovered and eliminated at the initial stage of acceptance and acceptance, and subsequent dynamic speed-up testing and trial operation are ensured to be safely and smoothly carried out. In the operation stage, in order to ensure good contact between the pantograph of the electric locomotive and the contact wire, reduce abrasion and avoid pantograph scraping or pantograph drilling accidents, the contact wire height and pull-out values need to be kept within a certain range, so that the contact wire height and pull-out values need to be subjected to normalized detection.
Contact detection and non-contact detection are generally adopted for automatic detection of the height of the overhead line system and the pull-out value. Compared with contact detection, the non-contact detection mode is more accurate, safe and intelligent. The non-contact detection method mainly comprises image detection and laser detection. The image detection is to install a CCD camera on the detection roof to acquire images and detect the leading height and the pulling value of the overhead line system by utilizing an image recognition technology. Due to different measurement references, the image detection method usually needs to compensate the vibration of the vehicle body, and the calculation process is complex. Meanwhile, the image detection method is easily influenced by factors such as weather and illumination, so that the problems of unstable detection quality, high target identification difficulty, low accuracy and the like are caused. At present, the traditional portable laser measuring equipment is generally adopted in the power supply section and the maintenance pipeline section in each work area site, static point measurement is carried out on the height and the pull-out value of a contact line in a skylight point, a laser distance meter is used for manually aiming and capturing the contact line at each positioning position of the contact line to carry out parameter sampling measurement, the labor intensity is high, the operation efficiency is low, the safety coefficient is low, and the detection function is single. Therefore, a new technology for detecting the height and the pull-out value of the contact network, which has the advantages of high detection efficiency, high detection precision, high automation degree and no contact, needs to be developed urgently to ensure the normal and stable operation of a contact network system.
The vehicle-mounted mobile laser scanning technology is a comprehensive measurement and detection technology integrating various sensors such as a Global Navigation Satellite System (GNSS), an Inertial Measurement Unit (IMU), a laser scanner, a digital camera, and a digital video camera on a mobile carrier. Various sensors automatically acquire various positions, postures, images and laser scanning data in a moving state, and non-contact spatial geographic information acquisition, processing and warehousing are realized through a unified geographic reference and data acquisition synchronization technology. In the operation process, the integrated three-dimensional laser scanning system is carried on a rail car, and the mass point cloud and the image data in the range of dozens of meters to hundreds of meters on two sides of the rail transit are rapidly collected through the movement of the carrier. And performing combined calculation on the ground GNSS base station, the mobile GNSS receiver, the ground control point, the IMU and the laser scanner data to obtain high-precision three-dimensional laser point cloud data. Compared with ground laser scanning and airborne laser scanning, the vehicle-mounted mobile scanning has a high-efficiency flexible data acquisition mode, is more and more applied to engineering practice, and the vehicle-mounted mobile scanning technology is one of the hotspots of current research. In recent years, more and more students utilize vehicle-mounted mobile laser point clouds to realize the research on the detection of the leading and the pulling values of the overhead line system.
Zhongsong of the southwest university of transportation discusses, in his master thesis, "a three-dimensional point cloud data-based catenary geometric parameter detection research," an RANSAC algorithm for improving random sampling consistency according to catenary point cloud characteristics, and extracting contact line point cloud data; and then, a conversion matrix from the camera coordinate system to the world coordinate system is obtained according to the space rotation translation information, and the lead height and the pull-out value of the contact line are calculated according to the conversion matrix. The method needs to convert the measuring reference from the center of the camera to the plane of the track, and the measuring result is easily influenced by factors such as vehicle body vibration, wheel track free space and the like.
The Chinese patent publication No. CN103852011A discloses a laser radar-based geometric parameter analysis method for a railway contact net, which utilizes information that the height leading value of a contact line point cloud is minimum, the extreme value of a pull-out value is located at an upright post of the contact net, and the like, so as to realize the detection of the geometric parameters of the contact net. However, this method has the following disadvantages: (1) under the condition that the pulling value of the contact net upright post is not an extreme value, the method is invalid; (2) this method does not detect lead-up and pull-out values at the location of the hanger suspension point.
Disclosure of Invention
In order to solve the problems in the prior art, the invention provides a method for automatically detecting the leading height and the pulling-out value of a contact network based on vehicle-mounted mobile laser point cloud, and the method has the characteristics of high efficiency, high accuracy, high automation degree and the like.
Therefore, the technical scheme of the invention is as follows:
a method for automatically detecting the leading height and the pull-out value of a contact network based on vehicle-mounted mobile laser point cloud comprises the following steps:
s1, point cloud data acquisition and pretreatment:
rapidly acquiring point cloud data of infrastructure and surrounding environment along the rail transit by using a vehicle-mounted mobile laser scanning system; based on laser point clouds collected along the line, laser point clouds comprising the rail and the contact network are segmented from the laser point clouds of the full section of the rail transit by setting range values of laser scanning emission angles corresponding to the rail and contact network point clouds and adopting an angle threshold segmentation algorithm; removing noise points in the segmented track and contact net laser point cloud by adopting a statistical filtering algorithm;
s2, steel rail point cloud segmentation and registration:
the improved least square fitting algorithm of the spatial straight line is utilized to carry out piecewise linear fitting on the track line to obtain equally spaced track line segments, and a cuboid segmentation algorithm is adopted to extract steel rail point cloud from the track laser point cloud after being pretreated by S1 based on the relatively stable spatial position relation between the track line segments subjected to piecewise fitting and the corresponding tracks; generating standard steel rail model point clouds according to a steel rail standard section diagram, and realizing accurate registration of the steel rail point clouds and the steel rail model point clouds by using a principal component analysis algorithm and a generalized iterative closest point algorithm, so as to obtain steel rail space geometric parameters, extract a rail top central line, calculate a line midpoint and further fit a line central line; updating space position parameters between the trajectory line and the steel rail point cloud by using the extracted rail top central line, and extracting and registering the next section of steel rail point cloud; iteratively repeating the steps to complete extraction and registration of all steel rail point clouds;
s3, automatically extracting point clouds of a contact line and a suspension point:
extracting contact lines and suspension points: extracting point clouds comprising contact lines and suspension points from the laser point clouds of the overhead line system obtained in S1 by adopting a cylindrical segmentation algorithm based on equal-interval track line segments obtained by segmentation linear fitting in S2 according to a relatively stable spatial position relation between the track lines and the overhead line system structure; performing three-dimensional linear model fitting on point clouds containing contact lines and suspension points by adopting a linear random sampling consistency algorithm, setting parameters of the random sampling consistency algorithm according to a cross section design drawing of a standard contact wire, and performing sectional extraction of the point clouds of the contact wire and automatic fitting of a three-dimensional linear model; in the area only containing the contact line, the dimensional characteristics of the point cloud are linear; at the suspension point location, the dimensional characteristics of the point cloud appear non-linear as the dropper and locator are connected to the contact line. Analyzing the dimensionality characteristics of the neighborhood of the target point by adopting a dimensionality analysis method, and setting a linear threshold value to extract a suspension point of the contact line;
s4, automatically detecting the height of the overhead line system and the pull-out value:
according to the definition and measurement method of the lead-up and the pull-out values, the detection results of the geometric parameters of the steel rail in S2 and the detection results of the contact line and the suspension point in S3 are used for automatically detecting the lead-up and the pull-out values of the contact line.
In the above step S1, the point cloud data collection and preprocessing includes the following steps:
(1) point cloud data acquisition:
along with the movement of the vehicle-mounted mobile laser scanning system, the laser scanner performs circumferential scanning in a direction perpendicular to the line direction or in a direction forming an angle of 45 degrees with the line direction to obtain laser point clouds in a traffic length range of a measured track and a certain distance range at two sides of the track; resolving the scanned data by utilizing the track line of the vehicle-mounted mobile laser scanning system to obtain laser point clouds under a full-line unified geographic space coordinate system, and exporting the laser point clouds under the full-line unified geographic space coordinate system into a universal data format;
(2) laser point cloud angle threshold segmentation:
dividing laser point clouds in a certain angle range according to angle information of rotation of a scanning head, an angle value of each stepping rotation of the scanning head and an angle starting point of circumferential scanning of the scanner, which are recorded in the data acquisition process of the scanner; according to the installation height, the installation angle and the geometric scale of the segmentation area of the scanner, calculating by utilizing a trigonometric function formula through space geometric analysis to obtain an angle range in which the segmentation area is located, or directly selecting a boundary point of a track or a contact net structure on the laser point cloud, and checking the scanning angle value of the boundary point; and then setting a floating value delta of a scanning angle value, traversing all laser point clouds according to a formula (1), and segmenting track and contact network point clouds:
Figure BDA0003220343880000041
in the formula:
Pi: the ith laser spot;
SegmentRailTrack、SegmentCatenary、SegmentOther: sequentially comprises a track point cloud, a contact net point cloud and other point clouds;
αmin、αmax: respectively a minimum scanning angle and a maximum scanning angle corresponding to the track, and the unit is degree;
βmin、βmax: respectively a minimum scanning angle and a maximum scanning angle corresponding to the contact network structure, and the unit is degree;
δ: dividing the floating value by an angle;
Figure BDA0003220343880000042
the scanning angle of the ith laser point in the track point cloud is represented by degree;
Figure BDA0003220343880000043
the scanning angle of the ith laser point in the contact network point cloud is measured in degrees;
preferably, the angle division floating value δ is 3 degrees;
(4) point cloud noise filtering:
and eliminating a small number of noise points in the point cloud data acquired by the three-dimensional laser scanning system by using a statistical filtering algorithm.
In the step (1), the general data format is an LAS format, the rotation angle range of the scanner is-180 to 180 degrees, the scanning starting direction is arranged right below the scanner, the scanning starting direction is 180 degrees when the scanner rotates right above the scanner, the left side angle of the advancing direction of the track line is a negative value, and the right side angle is a positive value. The certain distance range on the two sides of the track is the width range of 100 meters on the left side and the right side of the line respectively.
In the step (2), if the vehicle-mounted mobile laser scanning system comprises a plurality of scanners, laser point clouds are respectively divided according to the angles of the scanners, and then the laser point clouds divided by the scanners are combined.
Specifically, in step S2, the segmentation and registration steps of the steel rail point cloud are as follows:
(1) piecewise linear fitting of the trajectory: firstly, establishing a multi-dimensional space index structure of a rail transit laser point cloud; setting a Fixed step length (Fixed Depth), and performing piecewise linear fitting on a track line to obtain equidistant multi-segment lines;
(2) the steel rail point cloud segmentation and registration method comprises the following steps:
1) track line translation:
selecting a detection starting point, finding a track line segment AB corresponding to the starting point by using a distance discrimination method, and translating the track line segment AB to a right steel rail top surface point R of the starting point, wherein the translated line segment is A 'B';
2) cuboid segmentation:
establishing a cuboid bounding box by using the geometric dimension of the steel rail and the translated track line segment as A 'B', and dividing the steel rail point cloud; taking a straight line A 'B' as an axis, establishing vertical surfaces as the left and right side surfaces of the cuboid bounding box in the width of s on two sides within the length range of A 'B', wherein the width of s can be half of the width of the bottom surface of the steel rail; let the maximum elevation of A 'B' be HmaxMinimum height of HminThe height of the steel rail is Hrail(ii) a The steel rail top surface light band of the operation line can cause the dispersion of the point cloud of the steel rail top surface to be larger, so a height floating value epsilon is set, and the height of the cuboid bounding box top surface is determined as Hmax+ epsilon, bottom elevation Hmin-Hrail-epsilon, circularly traversing the track point cloud, and segmenting the steel rail point cloud in the cuboid bounding box; preferably, the height float value epsilon is 0.01 m;
3) point cloud registration:
establishing a gridding three-dimensional standard steel rail model according to a standard section diagram of a steel rail used for a detection line, and uniformly sampling the gridding three-dimensional standard steel rail model to form standard steel rail model point cloud; the classical ICP (iterative Closest point) registration algorithm and the improved algorithm thereof have relatively strict requirements on initial registration conditions, the degree of overlap of point clouds to be registered is required to be high, otherwise, the point clouds to be registered are easy to fall into a local optimal trap, and therefore, the coarse registration algorithm is required to be used for aligning the point clouds to be registered. After PCA coarse registration, the two groups of point clouds are basically overlapped, which provides a good initial condition for accurate registration. The precise registration of the steel rail point cloud is realized by adopting a GICP algorithm, the algorithm combines point-to-point ICP and point-to-surface ICP to a probability frame model, and a registration cost function is constructed by utilizing a covariance matrix of the point cloud, so that the speed and the robustness of the classical ICP registration are improved. Firstly, roughly aligning the steel rail point cloud to be registered with the standard steel rail model point cloud by using a rough registration algorithm, and then realizing the precise registration of the steel rail point cloud by using a GICP algorithm. Calculating the space geometric parameters of the steel rail by using the accurately registered steel rail model point cloud, calculating the line midpoints in a segmented manner on the basis of the reference rail, and sequentially connecting the line midpoints extracted in the segmented manner to form a line central line;
4) and (3) updating geometric parameters:
setting the center line of the point cloud top surface of the steel rail model after registration as A ' B ', replacing the R point with the B ' as a new detection starting point, and recalculating the geometric relationship of the trajectory line relative to the steel rail;
5) and (3) loop iteration:
and (4) circularly performing the steps 1) to 4), and using the new geometric parameters for extracting and registering the next section of steel rail point cloud until all the steel rail point clouds are extracted and registered.
And (5) segmenting and registering the point cloud of the left steel rail according to the steps 1) to 5).
Preferably, in the step (1), the multidimensional space index structure of the track traffic laser point cloud is a KD tree.
In the step 3), the registered three-dimensional geometric model of the steel rail is used for calculating geometric parameters of the steel rail to obtain track gauge and track direction information, the reference rail is determined according to the definition of the reference rail, then the reference rail is used as a basis, 1/2 standard track gauges (the standard track gauge defined in China is 1435mm) are horizontally deviated from another steel rail, the midpoint of the line is calculated in sections, the line midpoints extracted in sections are sequentially connected to form a line center line, and the formed line center line is subjected to smoothing treatment to obtain a relatively smooth line center line.
In step S3, the automatic extraction of the contact line and suspension point clouds includes the following steps:
(1) point cloud segmentation of contact lines and suspension points:
for the contact network structure after angle segmentation, a cylindrical segmentation algorithm is used for segmenting a contact line and suspension point cloud, firstly, a contact line point C is selected to obtain a three-dimensional coordinate of the contact line point C; secondly, acquiring a track line segment MN corresponding to the C point by adopting a distance discrimination method, and translating the track line segment MN to the C point, wherein the translated line segment is M 'N'; and finally, establishing a cylinder with the radius r by taking M 'N' as an axis, traversing the point cloud of the contact network structure, and segmenting to obtain the point cloud of the contact line and the suspension point. The contact lines are laid out in a zigzag pattern along the line direction, and the pull-out value is maximum at the positioner. TG/GD 124-2015 "high-speed railway contact network operation and maintenance rule" TG/GD 116-2017 "general speed railway contact network operation and maintenance rule" stipulate that the limit value of the contact line pulling value of the general speed railway and the high-speed railway is 450 mm. Considering the limit of the pull-out value, the cylinder radius r can be calculated according to equation (2):
Figure BDA0003220343880000061
in the formula: l is the distance between the contact net columns, and is generally 50 m;
preferably, the radius r is 0.1 m;
(2) extracting contact lines:
because the contact line has a small sectional area and is far away from a measurement system, when the scanning operation speed is too high, the contact line point cloud is rare, and the condition that only one point exists on the section of one contact line occurs, so that a linear model with random sampling consistency is adopted. The actual section diameter (height) of the contact line is set as a distance threshold value of linear sampling, and linear sampling is carried out to obtain contact line point cloud and a linear model starting and ending point M 'N'.
(3) Extracting suspension points:
solving the eigenvector and eigenvalue corresponding to the target point neighborhood point cloud set by Principal Component Analysis (PCA), and then calculating the eigenvalue lambda according to the eigenvalue lambda1、λ2、λ3(λ1>λ2>λ3) And judging the point cloud dimension characteristics. In the area only containing the contact line, the point cloud presents linear characteristics; at the suspension point, the dropper and the locator are connected with the contact line, and the dimensional characteristics of the point cloud are represented as nonlinearity. Traversing the point cloud after the cylindrical segmentation, calculating a linear characteristic value of the point cloud in the neighborhood of the radius r of the target point, setting a threshold L lambda of the linear characteristic value, and extracting suspension points. For the point cloud around the suspension point, the distance can be determined according to the distance λ1And judging the direction of the corresponding characteristic vector, wherein the point vertical to the ground is a hanging string hanging point, and the point approximately parallel to the hanging string hanging point is a positioning hanging point. Preferably, the linear eigenvalue threshold LλIs 0.9.
(4) And (3) loop iteration:
and (3) replacing the point C with the point N' of the linear model end point, and repeating the steps for 1-3 times until all contact lines and suspension points are extracted.
In step S4, the method for detecting the overhead contact system lead height and pull-out value is as follows:
(1) and (3) calculating the lead height:
extracting the bottom elevation of the contact line at the suspension point according to the division results of the suspension point and the contact line, finding the elevation of the rail surface at the corresponding position, and obtaining the height value of the contact line after calculating the difference;
(2) calculation of the pull-out value:
for a straight line section, the height of an outer rail does not exist, and the offset from the bottom point of the contact line at the positioning suspension point to the XY plane of the line center line is directly calculated; for the curve section, due to the presence of the outer rail superelevation, the pull-out value of the contact line is calculated according to equations (3) and (4):
a＝m+c (3)
Figure BDA0003220343880000071
in the formula: a: a catenary pull-out value;
m: positioning the horizontal distance between the contact line and the line center line;
c: positioning the horizontal distance between the center of the pantograph and the central line of the line;
h: the outer rail is ultrahigh;
h: a contact line height;
l: track gauge;
preferably, the method for searching the rail surface of the corresponding position is a distance discrimination method;
the spatial geometrical parameter calculation of the outer rail height according to the steel rail obtained in step S2 in claim 1.
The invention provides an automatic detection method for a contact net leading height and pulling value based on railway three-dimensional laser point cloud data acquired by a light vehicle-mounted mobile measurement system. Firstly, setting an angle segmentation threshold according to scanning angles corresponding to a track structure and a contact network structure, segmenting to obtain a track point cloud and a contact network point cloud, and removing point cloud noise by using a statistical filtering algorithm; secondly, based on the trajectory line of the mobile scanning, a cuboid segmentation algorithm is adopted to segment the steel rail point cloud, and a cylindrical segmentation algorithm is adopted to segment the contact line and the locating point cloud; thirdly, registering the steel rail point cloud by using PCA and GICP, extracting a contact line point cloud by using RANSAC, and extracting a suspension point cloud by using a dimensional analysis method; and finally, calculating the lead height and the pull-out value of the contact line according to the lead height and the pull-out value definitions.
The invention has the following beneficial effects:
(1) the existing detection methods are mostly in an image detection mode, and the image detection methods are easily influenced by factors such as weather and illumination, so that the problems of unstable detection quality, high target identification difficulty, low accuracy and the like are caused. The method is based on a laser scanning mode, is not influenced by factors such as illumination and the like, and has strong environmental adaptability and stable and reliable detection results.
(2) In the existing laser detection method, a laser scanner is carried on the top of a track, only contact line point clouds are obtained, a measurement reference is converted to a track plane from the center of the scanner according to the geometric relation between a track car and the track, and the calculation precision is easily influenced by factors such as car body vibration, wheel-track clearance and the like. The method obtains the geometric parameters of the track, the positions of the contact line and the suspension point based on the full-section point cloud of the railway line, directly calculates the contact line height conductance and the pull-out value, does not need to convert the measurement reference, and effectively improves the measurement precision.
(3) The existing steel rail point cloud segmentation algorithm is to utilize the characteristics of the maximum local height value of the surface of a steel rail or the direction change continuity of the point cloud on the surface of the steel rail and the like and utilize an elevation histogram statistical calculation method or a Kalman filtering method to classify the point cloud of the steel rail. The method is based on the stable position relation between the trajectory line of the mobile scanning and the steel rail, realizes the extraction of the point cloud by using a cuboid segmentation algorithm, realizes the accurate registration of the steel rail point cloud and the steel rail model point cloud by adopting PCA and GICP, and obtains the accurate space geometric parameters of the steel rail.
(4) The existing contact line extraction method is used for extracting contact line laser point clouds based on all laser point clouds of a contact network. The invention utilizes the cylindrical segmentation algorithm to obtain the contact line point cloud and the suspension point cloud by segmentation, and then carries out subsequent contact line extraction and suspension point extraction, thereby effectively improving the calculation efficiency.
(5) In the existing contact net height guiding and pulling value detection method based on laser point cloud, the position of a suspension point is not positioned, and the height guiding and pulling value of the position of the suspension point is concerned by a power department. According to the method, the position of a contact line suspension point is determined by adopting a dimension analysis method according to the neighborhood characteristics of the point cloud, and the type of the suspension point is identified according to the Z value component size of the eigenvector corresponding to the first characteristics of the neighborhood point cloud covariance matrix.
(6) The method provided by the invention has the advantages that the precision of measuring the contact wire height conductance and the pull-out value is 0-10 mm, and the measurement precision meets the requirements of the general technical specification of a high-speed railway power supply safety detection monitoring system (6C system).
(7) Compared with the traditional method for detecting the contact wire height and the pull-out value, the method provided by the invention not only improves the detection precision, but also can quickly realize the detection of the contact wire height and the pull-out value. And a vehicle-mounted mobile laser scanning mode is adopted, so that the on-line operation of detection personnel is avoided, and the safety operation risk is reduced. The invention can be used as a novel contact net lead height and pull-out value measuring mode, and can serve for the static acceptance of a newly-built railway line and the normalized detection of the existing railway line.
Drawings
FIG. 1 is a flow chart of a method for detecting the lead height and the pull-out value of a catenary of the invention;
FIG. 2 is a point cloud of a track and a catenary after pretreatment by the method of the present invention;
FIG. 3 is a schematic diagram of a steel rail point cloud segmentation according to the present invention;
FIG. 4 is a segmentation result of a steel rail point cloud in the present invention;
FIG. 5 is a registration flow diagram of the present invention;
FIG. 6a is a standard rail model;
FIG. 6b is a standard rail point cloud;
FIG. 7 is the registration result;
FIG. 8 shows the track detection results;
FIG. 9 is a schematic view of contact line extraction;
FIG. 10 is a schematic diagram of a dimensional analysis method;
FIG. 11 shows the contact line and hanging point test results;
FIG. 12 is a state diagram of point cloud data acquired using a light rail motion measurement system;
FIG. 13 is a schematic diagram of the high-speed rail route from salt city to Nantong;
FIG. 14 is a laser point cloud of a salt flux high-speed rail obtained in the present invention;
FIG. 15 is a graph comparing the results of the detection of lead-high and pull-out values with the data detected by a DJJ-8 meter according to the present invention.
Detailed Description
The technical scheme of the invention is further explained by combining the drawings and the specific embodiment.
The method for automatically detecting the leading height and the pulling value of the contact net comprises the following steps: point cloud data acquisition and pretreatment; partitioning and registering the point cloud of the steel rail; extracting point clouds of contact lines and suspension points, and calculating leading height and pulling values. As shown in fig. 1, the above steps are specifically as follows:
s1, data acquisition and preprocessing: rapidly acquiring data of infrastructure and surrounding environment along the rail transit by using a vehicle-mounted mobile laser scanning system; based on laser point clouds collected along the line, by setting range values of laser scanning emission angles corresponding to the track point clouds and the contact network point clouds, an angle threshold segmentation algorithm is adopted, the track and contact network laser point clouds are segmented from all the laser point clouds along the track traffic line, and then a statistical filtering algorithm is adopted to eliminate point cloud noise, and the specific method is as follows:
(1) point cloud data acquisition:
the laser scanner of the vehicle-mounted mobile laser scanning system adopts a mechanical rotary reflecting prism mode to realize 360-degree circumferential section scanning, and then the scanner carrying platform moves on the track to realize the scanning of a track traffic range and a certain distance range at two sides.
The scanner records the angle information of the rotation of the scanning head in the data acquisition process, the angle value of each stepping rotation of the scanning head is fixed, the angle starting point of each circular scanning is also fixed, generally, the circular scanning is finished by starting from the position right below the scanner and ending at the point. Because the contact net height guiding and pulling value only relates to the point cloud in the range of the track and the contact net, in order to reduce the influence of background point cloud and point cloud of other structures and improve the point cloud segmentation calculation speed and accuracy, the track and the contact net point cloud are obtained by segmentation according to the corresponding scanning angle of the track and the contact net structure.
And resolving the scanning data by utilizing the moving scanning track line to obtain laser point cloud under a full-line unified coordinate system. Exporting the preprocessed scanning data into a general data format, such as data in an LAS format, wherein the rotation angle range of a scanner recorded by the LAS-format laser point cloud data is-180 degrees, namely: the scanning starting direction is arranged right below the scanner, the scanning angle is 180 degrees when the scanner rotates to the position right above the scanner, the left angle of the advancing direction of the track line is a negative value, and the right angle is a positive value.
(2) Laser point cloud angle threshold segmentation:
dividing laser point clouds in a certain angle range according to angle information of the rotation of the scanning head, which can be recorded by the scanner in the data acquisition process, the angle value of each stepping rotation of the scanning head and the angle starting point of the circumferential scanning of the scanner; according to the parameters of the scanner such as the installation height, the installation angle, the geometric scale of the segmentation area and the like, calculating by utilizing a trigonometric function formula through a space geometric relationship to obtain an angle range where the segmentation area is located, or directly selecting a boundary point of a track or a contact net structure on the laser point cloud, and checking the scanning angle value of the selected point; in the vehicle-mounted scanning process, the change of the emission angle of the scanner caused by the shaking of the vehicle is considered, a floating value delta (such as 3-degree angle) of a certain angle value is set, all laser point clouds are traversed according to a formula (1), and the point cloud data of the track and the contact network are segmented out:
Figure BDA0003220343880000101
in the formula:
Pi: the ith laser spot;
SegmentRailTrack、SegmentCatenary、SegmentOther: sequentially comprises a track point cloud, a contact net point cloud and other point clouds;
αmin、αmax: respectively a minimum scanner angle and a maximum scanner angle corresponding to the track, and the unit is degree;
βmin、βmax: respectively a minimum scanner angle and a maximum scanner angle corresponding to the contact network structure, and the unit is degree;
Figure BDA0003220343880000102
the scanning angle of the ith laser point in the track point cloud is represented by degree;
Figure BDA0003220343880000103
the scanning angle of the ith laser point in the contact network point cloud is measured in degrees.
If the vehicle-mounted mobile laser scanning system comprises a plurality of scanners, laser point clouds are divided according to the angles of the scanners respectively, and then the laser point clouds divided by the scanners are combined.
(3) Point cloud denoising:
when the three-dimensional laser scanning system acquires point cloud data, the acquired point cloud contains a small amount of noise points due to various factors such as instrument oscillation, roughness of the surface of a measured object, mirror reflection, shielding of a shielding object and the like. The extraction of the steel rail point cloud and the contact line point cloud can be influenced by a small amount of noise points far away from the main structure, and the noise points are removed by using a statistical filtering algorithm.
S2, steel rail point cloud segmentation and registration:
the automatic segmentation of the laser point clouds of the left and right steel rails is realized by utilizing the relatively fixed spatial position relationship between the trajectory line of the vehicle-mounted mobile laser scanning system and the steel rail and adopting a data driving method.
And (3) carrying out laser point cloud segmentation through the angle segmentation threshold in the step (2), removing background point cloud data and other structure point clouds, and only keeping track and contact network point clouds. And (4) removing point cloud noise of the segmented track and contact net point cloud through the statistical filtering algorithm in the step (3), wherein the preprocessing result is shown in fig. 2.
The railway three-dimensional space structure and the scanning operation of the vehicle-mounted mobile laser scanning system have the characteristics that: (1) in the direction along the line, the two steel rails are basically parallel, and the distance between the inner side surfaces of the two steel rails is generally very close to the standard gauge (1435 mm); (2) the wheel clamp of the vehicle-mounted mobile laser scanning system runs along the rail on the rail, the relative position change of the scanner and the two rails is very small, and the line shape of the scanning track line is consistent with that of the railway line; (3) the steel rail is fixed on the sleeper or the ballastless track plate, and the steel rail is higher than the plane of the sleeper, the ballast and the track plate.
The moving laser scanning track file comprises spatial position and attitude data of the center of the laser scanning system. The space position data of the trajectory line is the space position points of the scanner under the reference coordinate system, and the space position points are connected in turn to form the trajectory line of the mobile scanning. When the whole scanning system is installed on a moving platform for scanning, the spatial position of the track line relative to the left and right rails is fixed (straight line) or continuously changes within a small range (curved line). The method utilizes the characteristics of relative fixation and full-line continuous change of the space positions of the track line and the steel rail and combines the inherent structural information of the steel rail to quickly realize the segmentation of the laser point cloud of the steel rail. Because the space geometric parameters and geometric information of the steel rail point cloud obtained by segmentation are unknown, the invention firstly establishes standard steel rail model point cloud (such as 60kg/m steel rail which is a common steel rail type in China) according to a standard section diagram of the steel rail used by a railway line, then adopts a mode of combining rough registration and precise registration to accurately register the segmented steel rail point cloud and the steel rail model point cloud, and obtains the space geometric parameters of the steel rail according to the registered steel rail model point cloud.
In order to obtain the space geometric parameters of the steel rail by segmentation, a segmentation schematic diagram is shown in figure 3, and the following two steps of work are required:
(1) and (3) piecewise fitting of a trajectory line:
in order to improve the extraction efficiency and the extraction precision of the steel rail point cloud, a segmentation method is adopted to extract the steel rail point cloud. And improving a least square fitting algorithm of the spatial straight line, and performing linear fitting with a fixed step length on the trajectory line to obtain a plurality of lines with the distance D. The principle of the improved least square fitting algorithm of the spatial straight line is as follows:
1) setting the trajectory line data set to have N points and the coordinate of the starting point as Pi(Xi,Yi,Zi) The coordinate of the jth point is Pj(Xj,Yj,Zj) The Euclidean distance between two points is dij。
2) If d isijIf not more than D, putting the j point into the set A. And (3) performing linear fitting on the points in the set A by adopting a least square fitting algorithm of a spatial straight line, wherein the principle is as follows:
the standard equation of the space straight line is
Figure BDA0003220343880000111
Performing equivalent transformation on the formula (2) to obtain a biplane equation of
Figure BDA0003220343880000112
Figure BDA0003220343880000121
In the formula (I), the compound is shown in the specification,
Figure BDA0003220343880000122
Figure BDA0003220343880000123
Figure BDA0003220343880000124
Figure BDA0003220343880000125
therefore, the spatial straight line can be regarded as obtained by intersecting two planes, and the two plane equations are fitted, so that the spatial straight line can be further obtained through fitting. Obtaining the optimal straight line parameter by using the residual sum of squares minimum theory, wherein the residual sum of squares of two planes is
Q1＝∑(xi-k1zi-b1)2 (5)
Q2＝∑(yi-k2zi-b2)2 (6)
Minimizing the sum of squares of the residuals, respectively for k1And k2Is derived by
Figure BDA0003220343880000126
Each of the formulas in the simultaneous formula (7) is solved to obtain the plane parameter of
Figure BDA0003220343880000127
Figure BDA0003220343880000128
Figure BDA0003220343880000129
Figure BDA00032203438800001210
3) Per PiPerpendicular to the straight line, with the foot being PNCalculating a point P with a distance D along the fitted lineM(in the scan direction), a straight line point P will be fittedMAs new PiAnd (4) point.
4) And (4) circulating steps 1), 2) and 3) until the piecewise linear fitting of the whole trajectory line is completed.
(2) The steel rail point cloud segmentation and registration method comprises the following steps:
the right rail will be described below as an example, and the process of the left rail is the same as that of the right rail.
1) Track line segment translation:
and selecting a right top surface central point R and a left top surface central point L of the steel rail, and obtaining a piecewise fitting track line segment AB corresponding to R and L by adopting a distance discrimination method. And translating the track line segment AB to the position R, wherein the translated track line segment is A 'B'.
2) Cuboid segmentation:
and establishing a cuboid bounding box by using the geometric dimension of the steel rail and the translated track line segment A 'B', and dividing the steel rail point cloud. The method specifically comprises the following steps: with the straight line A 'B' as an axis, the width of each side s (s can be half of the width of the bottom surface of the steel rail) within the length range of A 'B' establishes vertical surfaces as the left and right side surfaces of the cuboid bounding box. Let the maximum elevation of A 'B' be HmaxMinimum height of HminThe height of the steel rail is Hrail. The operating line steel rail light band can cause the dispersion degree of the point cloud of the top surface of the steel rail to be larger, so a height floating value epsilon is set, and the height of the top surface of the rectangular enclosure box is determined as Hmax+ epsilon, bottom elevation Hmin-Hrail- ε. And circularly traversing the track point cloud, segmenting the steel rail point cloud in the cuboid bounding box, and obtaining a steel rail point cloud segmentation result shown in figure 4.
3) Point cloud registration of steel rails:
after the segmentation of the steel rail point cloud is completed, the invention adopts a PCA and GICP registration mode to accurately register the steel rail point cloud and the steel rail model point cloud so as to obtain accurate space geometric parameters of the steel rail, and a registration flow chart is shown in figure 5, which specifically comprises the following steps:
a) generating a standard steel rail model point cloud:
according to a standard section diagram of the steel rail type used by the detection line, a gridding three-dimensional steel rail model is established and is uniformly sampled to form standard steel rail model point cloud, as shown in fig. 6a and 6 b.
b) PCA-based point cloud coarse registration
The classical icp (iterative closed point) registration algorithm and its improved algorithm both have relatively strict requirements on initial registration conditions, require high overlap of point clouds to be registered, otherwise are liable to sink into a local optimal trap, so that the point clouds to be registered need to be roughly aligned by using a coarse registration algorithm. Because the length of the steel rail of the railway is far greater than the height and the width of the steel rail, the space of the point cloud of the steel rail presents linear distribution, and the Principal Component Analysis (PCA) algorithm is applied to the rough registration of the point cloud of the steel rail. The specific procedure for coarse registration by PCA is as follows:
assuming that the extracted point cloud of the steel rail is Si(xi,yi,zi) I is 1,2,3 …, m; the point cloud of the steel rail model is Gi(xi,yi,zi) I is 1,2,3 …, n. Calculating the mean of two groups of point clouds
Figure BDA0003220343880000131
And
Figure BDA0003220343880000132
Figure BDA0003220343880000133
Figure BDA0003220343880000134
covariance matrix COV for respectively constructing two groups of point cloudsSAnd COVG。
Figure BDA0003220343880000135
Figure BDA0003220343880000136
Computing a covariance matrix COV by a singular value decomposition algorithm (SVD)SAnd COVGThe eigenvalues and eigenvectors.
Rigid transformation matrix Trans of PCA registration is composed of rotation matrix RPAnd translation matrix TPAnd (4) forming.
Figure BDA0003220343880000141
Figure BDA0003220343880000142
Figure BDA0003220343880000143
In the formula of UGAnd USRepresenting a matrix consisting of covariance matrix eigenvectors.
After the rigid transformation matrix Trans conversion calculated by PCA, the two groups of point clouds are basically overlapped, which provides a good initial condition for accurate registration. The invention adopts GICP algorithm to carry out precise registration, the algorithm combines point-to-point ICP and point-to-surface ICP to a probability frame model, a registration cost function is constructed by utilizing a covariance matrix of point cloud, the speed and the robustness of classical ICP registration are improved, and the principle is as follows:
suppose two sets of point clouds a and B are found by nearest point search:
Figure BDA0003220343880000144
in the formula, aiAnd biAre matched with each other.
In the probabilistic model, we assume that there is a set of potential points
Figure BDA0003220343880000145
And
Figure BDA0003220343880000146
(scrambling code):
Figure BDA0003220343880000147
two sets of point clouds a and B may be generated by gaussian distributions:
Figure BDA0003220343880000148
in the formula: ciIs the covariance matrix of each point.
Defining the rigid body translation rotation matrix as T, and the registration error d of each pair of corresponding pointsiThen it is:
di＝bi-Tai (22)
because a isiAnd biIndependent of each other and all obey a gaussian distribution, so d also obeys a gaussian distribution:
Figure BDA0003220343880000149
utilizing maximum likelihood estimation, calculating a cost function of the rigid translation rotation matrix T as:
Figure BDA0003220343880000151
the transformation matrix obtained by PCA and GICP calculation is used for realizing the accurate registration of the steel rail point cloud and the steel rail model point cloud, and the registration result is shown in FIG. 7. Calculating the geometric parameters of the track by using the registered steel rail model point cloud to obtain information such as actual track gauge, track direction and the like, determining a reference track according to the definition of the reference track, horizontally shifting 1/2 standard track gauge to another steel rail on the basis of the reference track (the standard track gauge defined in China is 1435mm), calculating the line midpoints in sections, sequentially connecting the line midpoints extracted by the sections to form a line central line, and preferably, smoothing the formed line central line to obtain a relatively smooth line central line, as shown in fig. 8.
4) And (3) updating geometric parameters:
and (4) setting the center line of the point cloud top surface of the steel rail model after registration as A ' B ', replacing the R point with the B ' as a new detection starting point, and recalculating the geometric relation of the trajectory line relative to the steel rail.
5) And (3) loop iteration:
and (4) circularly performing the steps 1) to 4), and using the new geometric parameters for extracting and registering the next section of steel rail point cloud until all the steel rail point clouds are extracted and registered.
S3, extracting point clouds of the contact line and the suspension point:
the contact lines of the Chinese electrified railway are distributed in a zigzag shape, the pulling values of the contact lines continuously change along the line, and the extreme values are positioned at the positioner of the contact net upright post. The method comprises the following steps of utilizing a local relatively fixed spatial position relation between a track line and a contact line of vehicle-mounted mobile scanning, using a cylindrical segmentation algorithm to segment the contact line and suspension point clouds, using a segmentation schematic diagram shown in figure 9, extracting the contact line point clouds by adopting linear random sampling consistency, and extracting the suspension points by adopting a dimension analysis method, wherein the specific flow is as follows:
(1) segmentation of contact line and suspension point clouds:
selecting a contact line point C for the contact line point cloud subjected to angle segmentation and statistical filtering denoising to obtain a three-dimensional coordinate of the contact line point; secondly, acquiring a track line segment MN corresponding to the C point by adopting a distance discrimination method, and translating the track line segment MN to the C point, wherein the translated line segment is M 'N'; and finally, establishing a cylinder with the radius r by taking M 'N' as an axis, traversing the point cloud of the contact network structure, and segmenting to obtain the point cloud of the contact line and the suspension point. The contact lines are laid out in a zigzag pattern along the line direction, and the pull-out value is maximum at the positioner. TG/GD 124-2015 "high-speed railway contact network operation and maintenance rule" TG/GD 116-2017 "general speed railway contact network operation and maintenance rule" stipulate that the limit value of the contact line pulling value of the general speed railway and the high-speed railway is 450 mm. The cylinder radius r can be calculated according to equation (25) in consideration of the limit of the pull-out value.
Figure BDA0003220343880000152
In the formula: and L is the distance between the contact net columns, and is generally 50 m.
(2) Extracting contact lines:
because the contact line has a small sectional area and is far away from the scanner, when the scanning operation speed is too high, the contact line point cloud is rare, and the condition that only one point exists on the section of one contact line occurs, so that a linear model with random sampling consistency is adopted. The actual section diameter (height) of the contact line is set as a distance threshold value of linear sampling, and linear sampling is carried out to obtain contact line point cloud and a linear model starting and ending point M 'N'.
(3) Extracting suspension points:
solving the eigenvector and eigenvalue corresponding to the target point neighborhood point cloud set by Principal Component Analysis (PCA), and then calculating the eigenvalue lambda according to the eigenvalue lambda1、λ2、λ3(λ1>λ2>λ3) And judging the point cloud dimension characteristics, and showing a figure 10 by the principle of a dimension analysis method. In the area only containing the contact line, the point cloud presents linear characteristics; at the suspension point location, the neighborhood point cloud is non-linear due to the dropper or locator and contact line connection. Traversing the point cloud after the cylindrical segmentation, and calculating the linear characteristic value L of the point cloud in the neighborhood of the radius r of the target pointλSetting a linear characteristic value threshold value, and extracting suspension points. For the suspension point location point cloud, the method can be based on the distance λ1And judging the direction of the corresponding characteristic vector, wherein the point vertical to the ground is a hanging string hanging point, and the point approximately parallel to the hanging string hanging point is a positioning hanging point.
Preferably, the radius r is 0.1m, and the threshold value L of the linear eigenvalue isλIs 0.9.
(4) And (3) loop iteration:
and (3) replacing the point C with the point N' of the linear model end point, taking the point C as a new processing starting point, repeating the process until the extraction of all contact lines and suspension points is completed, and obtaining an extraction result shown in figure 11.
S4, automatically detecting the height of the overhead line system and the pull-out value:
according to the related definition of the lead height and the pull-out value, the measurement of the lead height and the pull-out value can be realized by utilizing the detection results of the track, the contact line and the suspension point, and the specific method comprises the following steps:
(1) and (3) calculating the lead height:
and extracting the bottom elevation of the contact line at the suspension point according to the division results of the suspension point and the contact line, finding the elevation of the rail surface at the corresponding position, and obtaining the height value of the contact line after calculating the difference.
(2) Calculation of the pull-out value:
for a straight line section, no outer rail height exists, and the offset of the bottom point of the contact line at the suspension point to the XY plane of the line center line is directly calculated. For the curve section, the pull-out value of the contact line is calculated as equations (26) and (27) due to the presence of the outer rail superelevation.
a＝m+c (26)
Figure BDA0003220343880000161
In the formula: a: a catenary pull-out value;
m: positioning the horizontal distance between the contact line and the line center line;
c: positioning the horizontal distance between the center of the pantograph and the central line of the line;
h: the outer rail is ultrahigh;
h: a contact line height;
l: and (4) track gauge.
Example 1
Step 1, point cloud data acquisition and pretreatment: to verify the correctness of the method of the present invention, point cloud data was acquired using a light railway mobile measurement system, see fig. 12. The light railway mobile measuring system consists of a light assembled rail car, a positioning system (comprising a GNSS receiver and an inertial measuring unit), a laser scanner and a panoramic camera, has the characteristics of convenience in online, flexible operation mode, high detection precision and detection efficiency and the like, and can acquire railway three-dimensional laser point cloud and high-definition panoramic image data. The highest measuring speed of the light railway mobile measuring system is 30km/h, the laser emission frequency is 1000kHz, the scanning line is 200 lines/second, the maximum measuring range is 119m, and the relative measuring precision is better than 1 mm.
In order to evaluate the reliability and precision of the method, a scanning experiment is selected to be carried out on a newly-built high-speed rail. The high-speed rail for salt communication is a high-speed railway connecting salt cities in Jiangsu province and Nantong cities, is an important component of a coastal channel which is one of main channels of eight longitudinal and eight transverse high-speed railways in 'Medium-long term railway network planning' (2016 revised edition), and comprises a salt city station to a Nantong west station, wherein the total length of a main line is 157.098 kilometers, 6 stations are arranged, the maximum design speed is 350 kilometers per hour, and the position of the line is shown in figure 13. Scanning operation is carried out on the whole line of the salt-passing high-speed rail at the speed of 20km/h by using a light railway mobile measuring system, and high-precision and high-density three-dimensional point cloud data of the salt-passing high-speed rail line are collected, wherein the average point density of the track is 9000 points/m2The mean point density of the contact lines was 60 points/m, and the acquired haloton high-speed rail point cloud data is shown in fig. 14.
The method comprises the steps of obtaining three-dimensional laser point clouds along the rail transit line through a light railway mobile measurement system, and achieving laser point cloud segmentation of the rail and the contact net by setting angle thresholds of the rail and the contact net and adopting an angle segmentation algorithm. And removing point cloud noise of the steel rail and the contact net point cloud obtained by segmentation by using a statistical filtering algorithm. Compared with all three-dimensional laser point cloud data, the preprocessed rail and contact net laser point cloud is more beneficial to subsequent processing work.
Step 2: steel rail point cloud segmentation and registration:
and fitting the trajectory line at intervals of 1m in the direction of the railway line, taking the fitted straight line segment as a reference line, and segmenting by adopting a cuboid segmentation algorithm by utilizing the relatively fixed spatial position relation between the mobile scanning trajectory line and the rail to realize the segmentation of the laser point cloud of the steel rail. And the method of PCA coarse registration and GICP fine registration is adopted to realize the accurate registration of the steel rail point cloud and the steel rail model point cloud, the accurate measurement of the steel rail point cloud space geometric parameters is realized based on the conversion relation between the registered steel rail model point cloud parameters and the registration, and the line center line is further fitted.
And step 3: extraction of contact lines and suspension points:
and (3) obtaining a contact line and suspension point cloud by utilizing the relatively stable position relation between the track line and the contact network structure which are subjected to segmented fitting in the step (2) and adopting a cylindrical segmentation algorithm. And extracting the contact line point cloud by adopting a linear random sampling consistency algorithm for the contact line point cloud and the suspension point cloud after segmentation, and extracting the suspension point cloud by using a dimension analysis method.
And 4, step 4: calculating the leading height and the pull-out value:
according to the definition and the measuring method of the lead height and the pull-out value, the accurate measurement of the lead height and the pull-out value is realized by utilizing the detection results of the track, the contact line and the suspension point.
And (3) analyzing the experimental precision:
the DJJ-8 laser contact net detector developed by Jinan blue-moving laser technology Limited of Shandong academy of sciences, Shandong province is widely used and accepted in domestic static detection of the guide height and the pull-out value of a contact net, the guide height of the detector is +/-3 mm, the pull-out value is +/-4 mm, the measurement result of the detection method is compared with the detection data of a DJJ-8 measuring instrument, and the comparison result is shown in figure 15. Through statistics, the median error of the difference of the pull-out values is +/-6.6 mm, and the maximum difference is 9.9 mm; the mean error of the height guiding difference is +/-6.3 mm, the maximum difference is 9.6mm, and the measurement precision meets the requirements of the general technical specification of a high-speed railway power supply safety detection monitoring system (6C system).

Claims (10)

1. A method for automatically detecting the leading height and the pull-out value of a contact network based on vehicle-mounted mobile laser point cloud comprises the following steps:
s1, point cloud data acquisition and pretreatment:
rapidly acquiring point cloud data of infrastructure and surrounding environment along the rail transit by using a vehicle-mounted mobile laser scanning system; based on laser point clouds collected along the line, laser point clouds comprising the rail and the contact network are segmented from the laser point clouds of the full section of the rail transit by setting range values of laser scanning emission angles corresponding to the rail and contact network point clouds and adopting an angle threshold segmentation algorithm; removing noise points in the segmented track and contact net laser point cloud by adopting a statistical filtering algorithm;
s2, steel rail point cloud segmentation and registration:
the improved least square fitting algorithm of the spatial straight line is utilized to carry out piecewise linear fitting on the track line to obtain equally spaced track line segments, and a cuboid segmentation algorithm is adopted to extract steel rail point cloud from the track laser point cloud obtained by preprocessing of S1 based on the relatively stable spatial position relation between the track line segments subjected to piecewise fitting and the corresponding tracks; generating standard steel rail model point clouds according to a steel rail standard section diagram, and realizing accurate registration of the steel rail point clouds and the steel rail model point clouds by using a principal component analysis algorithm and a generalized iterative closest point algorithm, so as to obtain steel rail space geometric parameters, extract a rail top central line, calculate a line midpoint and further fit a line central line; updating space position parameters between the trajectory line and the steel rail point cloud by using the extracted rail top central line, and extracting and registering the next section of steel rail point cloud; iteratively repeating the steps to complete extraction and registration of all steel rail point clouds;
s3, automatically extracting point clouds of a contact line and a suspension point:
extracting contact lines and suspension points: extracting point clouds comprising contact lines and suspension points from the laser point clouds of the overhead line system obtained in S1 by adopting a cylindrical segmentation algorithm based on equal-interval track line segments obtained by segmentation linear fitting in S2 according to a relatively stable spatial position relation between the track lines and the overhead line system structure; performing three-dimensional linear model fitting on point clouds containing contact lines and suspension points by adopting a linear random sampling consistency algorithm, setting parameters of the random sampling consistency algorithm according to a cross section design drawing of a standard contact wire, and performing sectional extraction of the point clouds of the contact wire and automatic fitting of a three-dimensional linear model; analyzing the dimensionality characteristics of the neighborhood of the target point by adopting a dimensionality analysis method, and setting a linear threshold value to extract a suspension point of the contact line;
s4, automatically detecting the height of the overhead line system and the pull-out value:
according to the definition and measurement method of the lead-up and the pull-out values, the detection results of the geometric parameters of the steel rail in S2 and the detection results of the contact line and the suspension point in S3 are used for automatically detecting the lead-up and the pull-out values of the contact line.
2. The method for automatically detecting the leading height and the pulling-out value of the overhead line system of claim 1, wherein the point cloud data acquisition and pretreatment in the step S1 comprises the following steps:
(1) point cloud data acquisition:
along with the movement of the vehicle-mounted mobile laser scanning system, the laser scanner performs circumferential scanning in a direction perpendicular to the line direction or in a direction forming an angle of 45 degrees with the line direction to obtain laser point clouds in a traffic length range of a measured track and a certain distance range at two sides of the track; resolving the scanned data by utilizing the track line of the vehicle-mounted mobile laser scanning system to obtain laser point clouds under a full-line unified geographic space coordinate system, and exporting the laser point clouds under the full-line unified geographic space coordinate system into a universal data format;
(2) laser point cloud angle threshold segmentation:
dividing laser point clouds in a certain angle range according to angle information of rotation of a scanning head, an angle value of each stepping rotation of the scanning head and an angle starting point of circumferential scanning of the scanner, which are recorded in the data acquisition process of the scanner; according to the installation height, the installation angle and the geometric scale of the segmentation area of the scanner, calculating by utilizing a trigonometric function formula through space geometric analysis to obtain an angle range in which the segmentation area is located, or directly selecting a boundary point of a track or a contact net structure on the laser point cloud, and checking the scanning angle value of the boundary point; and then setting a floating value delta of a scanning angle value, traversing all laser point clouds according to a formula (1), and segmenting track and contact network point clouds:
Figure FDA0003220343870000021
in the formula:
Pi: the ith laser spot;
SegmentRailTrack、SegmentCatenary、SegmentOther: sequentially comprises a track point cloud, a contact net point cloud and other point clouds;
αmin、αmax: respectively corresponding to the trackA scanning angle and a maximum scanning angle, the unit being degree;
βmin、βmax: respectively a minimum scanning angle and a maximum scanning angle corresponding to the contact network structure, and the unit is degree;
δ: dividing the floating value by an angle;
Figure FDA0003220343870000022
the scanning angle of the ith laser point in the track point cloud is represented by degree;
Figure FDA0003220343870000023
the scanning angle of the ith laser point in the contact network point cloud is measured in degrees;
preferably, the angle division floating value δ is 3 degrees;
(3) point cloud noise filtering:
and eliminating a small number of noise points in the point cloud data acquired by the three-dimensional laser scanning system by using a statistical filtering algorithm.
3. The method for automatically detecting the leading height and the pulling-out value of the overhead line system of claim 2, which is characterized in that: in the step (1), the general data format is an LAS format, the rotation angle range of the scanner is-180 to 180 degrees, the scanning starting direction is arranged right below the scanner, the scanning starting direction is 180 degrees when the scanner rotates right above the scanner, the left side angle of the advancing direction of the track line is a negative value, and the right side angle is a positive value.
4. The method for automatically detecting the contact line lead height and the pull-out value according to claim 2, wherein: in the step (2), if the vehicle-mounted mobile laser scanning system comprises a plurality of scanners, laser point clouds are respectively divided according to the angles of the scanners, and then the laser point clouds divided by the scanners are combined.
5. The method for automatically detecting the leading height and the pulling-out value of the overhead line system of claim 2, which is characterized in that: in the step (1), the certain distance range of the two sides of the track is the width range of 100 meters on the left side and the right side of the line respectively;
6. the method for automatically detecting the leading height and the pulling-out value of the overhead line system of claim 1, which is characterized in that: the segmentation and registration steps of the steel rail point cloud in the step S2 are as follows:
(1) piecewise linear fitting of the trajectory: firstly, establishing a multi-dimensional space index structure of a rail transit laser point cloud; setting a Fixed step length (Fixed Depth), and performing piecewise linear fitting on a track line to obtain equidistant multi-segment lines;
(2) the steel rail point cloud segmentation and registration method comprises the following steps:
1) track line translation:
selecting a detection starting point, finding a track line segment AB corresponding to the starting point by using a distance discrimination method, translating the track line segment AB to a right steel rail top surface point R of the starting point, wherein the translated line segment is A 'B',
2) cuboid segmentation:
establishing a cuboid bounding box by using the geometric dimension of the steel rail and the translated track line segment as A 'B', dividing the steel rail point cloud, establishing vertical surfaces as the left side surface and the right side surface of the cuboid bounding box by taking the straight line A 'B' as an axis and the widths of s on two sides in the length range of the A 'B', wherein the width of s can be half of the width of the bottom surface of the steel rail; let the maximum elevation of A 'B' be HmaxMinimum height of HminThe height of the steel rail is Hrail(ii) a Setting a height floating value epsilon, and determining the height of the top surface of the rectangular enclosure box as Hmax+ epsilon, bottom elevation Hmin-Hrail-epsilon, circularly traversing the track point cloud, and segmenting the steel rail point cloud in the cuboid bounding box; preferably, the height float value epsilon is 0.01 m;
3) point cloud registration:
establishing a gridding three-dimensional standard steel rail model according to a standard section diagram of a steel rail used for a detection line, and uniformly sampling the gridding three-dimensional standard steel rail model to form standard steel rail model point cloud; firstly, roughly aligning steel rail point clouds to be registered with standard steel rail model point clouds by using a rough registration algorithm, then realizing precise registration of the steel rail point clouds by using a GICP algorithm, calculating space geometric parameters of a steel rail by using the steel rail model point clouds after precise registration, calculating line midpoints in sections on the basis of a reference track, and sequentially connecting the line midpoints extracted in sections to form a line central line;
4) and (3) updating geometric parameters:
setting the center line of the point cloud top surface of the steel rail model after registration as A ' B ', replacing the R point with the B ' as a new detection starting point, and recalculating the geometric relationship of the trajectory line relative to the steel rail;
5) and (3) loop iteration:
circularly performing the steps 1) to 4), and using the new geometric parameters for extracting and registering the next section of steel rail point cloud until all the steel rail point clouds are extracted and registered;
and (5) segmenting and registering the point cloud of the left steel rail according to the steps 1) to 5).
7. The method for automatically detecting the leading height and the pulling-out value of the overhead line system of claim 6, which is characterized in that: in the step (1), the multidimensional space index structure of the track traffic laser point cloud is a KD tree.
8. The method for automatically detecting the leading height and the pulling-out value of the overhead line system of claim 6, which is characterized in that: in step 3), the registered three-dimensional geometric model of the steel rail is used for calculating geometric parameters of the steel rail to obtain track gauge and track direction information, after a reference track is determined according to the definition of the reference track, the reference track is used as a basis, 1/2 standard track gauges are horizontally deviated from another steel rail, the midpoint of a line is calculated in sections, the midpoint of the line extracted by the sections is sequentially connected to form a line center line, and the formed line center line is subjected to smoothing treatment to obtain a relatively smooth line center line.
9. The method for automatically detecting the leading height and the pulling-out value of the overhead line system of claim 1, which is characterized in that: the automatic extraction of the contact line and suspension point clouds in step S3 includes the following steps:
(1) point cloud segmentation of contact lines and suspension points:
for the contact network structure after angle segmentation, a cylindrical segmentation algorithm is used for segmenting a contact line and suspension point cloud, firstly, a contact line point C is selected to obtain a three-dimensional coordinate of the contact line point C; secondly, acquiring a track line segment MN corresponding to the C point by adopting a distance discrimination method, and translating the track line segment MN to the C point, wherein the translated line segment is M 'N'; finally, establishing a cylinder with the radius r by taking M 'N' as an axis, traversing the point cloud of the contact network structure, and segmenting to obtain a point cloud of a contact line and a suspension point; the contact lines are arranged in a zigzag shape along the line direction, and the pull-out value is maximum at the positioner; the limit value of the contact line pulling value of the ordinary speed railway and the high speed railway is 450mm, and the radius r of the cylinder can be calculated according to the formula (2) according to the limit condition of the pulling value:
Figure FDA0003220343870000041
in the formula: l is the distance between the contact net columns, and is generally 50 m;
preferably, the radius r is 0.1 m;
(2) extracting contact lines:
because the contact line has a small sectional area and is far away from a measurement system, when the scanning operation speed is too high, the contact line point cloud is rare, and the condition that only one point exists on the section of one contact line occurs, so a linear model with random sampling consistency is adopted; setting the actual section diameter (height) of the contact line as a distance threshold value of linear sampling, and linearly sampling to obtain a contact line point cloud and a linear model starting and ending point M 'N';
(3) extracting suspension points:
solving the eigenvector and eigenvalue corresponding to the target point neighborhood point cloud set by Principal Component Analysis (PCA), and then calculating the eigenvalue lambda according to the eigenvalue lambda1、λ2、λ3(λ1>λ2>λ3) Judging the point cloud dimension characteristics; in the area only containing the contact line, the point cloud presents linear characteristics; at the suspension point, the dropper and the locator are connected with the contact line, and the dimensional characteristics of the point cloud are represented as nonlinearity; traversing cylinderCalculating the linear characteristic value of the point cloud in the neighborhood of the radius r of the target point of the cut point cloud, and setting a threshold value L of the linear characteristic valueλExtracting suspension points; for the point cloud around the suspension point, the distance can be determined according to the distance λ1Judging the direction of the corresponding characteristic vector, wherein the point vertical to the ground is a hanger suspension point, the point approximately parallel to the ground is a positioning suspension point, and preferably, the linear characteristic value threshold value L isλIs 0.9;
(4) and (3) loop iteration:
and (3) replacing the point C with the point N' of the linear model end point, taking the point C as a new detection starting point, and repeating the steps (1) to (3) until all contact lines and suspension points are extracted.
10. The method for automatically detecting the leading height and the pulling-out value of the overhead line system of claim 1, which is characterized in that: the method for detecting the height of the overhead line system and the pull-out value in the step S4 is as follows:
(1) and (3) calculating the lead height:
extracting the bottom elevation of the contact line at the suspension point according to the division results of the suspension point and the contact line, finding the elevation of the rail surface at the corresponding position, and obtaining the height value of the contact line after calculating the difference;
(2) calculation of the pull-out value:
for a straight line section, the height of an outer rail does not exist, and the offset from the bottom point of the contact line at the positioning suspension point to the XY plane of the line center line is directly calculated; for the curve section, due to the presence of the outer rail superelevation, the pull-out value of the contact line is calculated according to equations (3) and (4):
a＝m+c (3)
Figure FDA0003220343870000051
in the formula: a: a catenary pull-out value;
m: positioning the horizontal distance between the contact line and the line center line;
c: positioning the horizontal distance between the center of the pantograph and the central line of the line;
h: the outer rail is ultrahigh;
h: a contact line height;
l: the track gauge is arranged on the track base,
preferably, the method for searching the rail surface of the corresponding position is a distance discrimination method;
and calculating the spatial geometric parameters of the steel rail obtained in the outer rail ultrahigh step S2.
CN202110955517.6A 2021-08-19 2021-08-19 Automatic detection method for contact net leading height and pulling value based on vehicle-mounted mobile laser point cloud Active CN113870123B (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN202110955517.6A CN113870123B (en) 2021-08-19 2021-08-19 Automatic detection method for contact net leading height and pulling value based on vehicle-mounted mobile laser point cloud
PCT/CN2021/124110 WO2023019709A1 (en) 2021-08-19 2021-10-15 Automatic detection method of conductor height and pull-out value of overhead line system based on vehicle-mounted mobile laser point cloud

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110955517.6A CN113870123B (en) 2021-08-19 2021-08-19 Automatic detection method for contact net leading height and pulling value based on vehicle-mounted mobile laser point cloud

Publications (2)

Publication Number Publication Date
CN113870123A true CN113870123A (en) 2021-12-31
CN113870123B CN113870123B (en) 2023-01-03

Family

ID=78990708

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110955517.6A Active CN113870123B (en) 2021-08-19 2021-08-19 Automatic detection method for contact net leading height and pulling value based on vehicle-mounted mobile laser point cloud

Country Status (2)

Country Link
CN (1) CN113870123B (en)
WO (1) WO2023019709A1 (en)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114440767A (en) * 2022-01-26 2022-05-06 株洲时代电子技术有限公司 Railway comprehensive inspection method
CN114782925A (en) * 2022-06-17 2022-07-22 四川省公路规划勘察设计研究院有限公司 Vehicle-mounted LIDAR data-based highway guardrail vectorization method and device
CN115546266A (en) * 2022-11-24 2022-12-30 中国铁路设计集团有限公司 Multi-strip airborne laser point cloud registration method based on local normal correlation
CN115657067A (en) * 2022-12-26 2023-01-31 中国铁路设计集团有限公司 Multi-scale array type laser radar measuring system and method for acquiring track boundary conditions
CN115824070A (en) * 2023-02-23 2023-03-21 湖南睿图智能科技有限公司 Method and device for measuring size of steel rail piece
CN116128835A (en) * 2023-02-02 2023-05-16 南方电网数字电网研究院有限公司 Point cloud analysis-based power transmission wire sag measurement method, device and equipment
CN116486012A (en) * 2023-04-27 2023-07-25 中国民用航空总局第二研究所 Aircraft three-dimensional model construction method, storage medium and electronic equipment
CN116862976A (en) * 2023-06-08 2023-10-10 中铁第四勘察设计院集团有限公司 Rail center line extraction method and system based on unmanned aerial vehicle laser point cloud

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116385684B (en) * 2023-04-10 2023-11-03 长江水利委员会水文局 Riverway section automatic extraction method based on reservoir point cloud
CN116147525B (en) * 2023-04-17 2023-07-04 南京理工大学 Pantograph contour detection method and system based on improved ICP algorithm
CN116452467B (en) * 2023-06-16 2023-09-22 山东曙岳车辆有限公司 Container real-time positioning method based on laser data
CN116543037B (en) * 2023-07-05 2024-03-26 中国铁路设计集团有限公司 Method for extracting center of track bearing platform of CRTSIII ballastless track plate
CN116892912B (en) * 2023-09-11 2023-12-22 中国铁路设计集团有限公司 Rapid accurate measurement and accurate adjustment method for track bearing platform of CRTSIII ballastless track plate
CN117162098B (en) * 2023-10-07 2024-05-03 合肥市普适数孪科技有限公司 Autonomous planning system and method for robot gesture in narrow space
CN117576144B (en) * 2024-01-15 2024-03-29 湖北工业大学 Laser point cloud power line extraction method and device and electronic equipment

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2015039375A1 (en) * 2013-09-17 2015-03-26 中国科学院深圳先进技术研究院 Method and system for automatically optimizing quality of point cloud data
CN110647798A (en) * 2019-08-05 2020-01-03 中国铁路设计集团有限公司 Automatic track center line detection method based on vehicle-mounted mobile laser point cloud
CN112414309A (en) * 2020-11-25 2021-02-26 北京交通大学 High-speed rail contact line height-guiding and pull-out value inspection method based on airborne laser radar
CA3163192A1 (en) * 2019-12-17 2021-06-24 Telefonos, Lineas Y Centrales, S.A. Method for in-situ and real-time collection and processing of geometric parameters of railway lines

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103852011A (en) * 2014-03-20 2014-06-11 北京天格高通科技有限公司 Railway overhead line system geometric parameter analysis method based on laser radar
EP3248140A4 (en) * 2015-01-20 2018-12-05 Solfice Research, Inc. Real time machine vision and point-cloud analysis for remote sensing and vehicle control
CN108564575B (en) * 2018-04-11 2022-03-15 西南交通大学 Non-contact type contact network parameter detection method based on three-dimensional point cloud data
AT520291B1 (en) * 2018-06-01 2019-03-15 Plasser & Theurer Export Von Bahnbaumaschinen Gmbh Method for determining an actual position of rails of a track

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2015039375A1 (en) * 2013-09-17 2015-03-26 中国科学院深圳先进技术研究院 Method and system for automatically optimizing quality of point cloud data
CN110647798A (en) * 2019-08-05 2020-01-03 中国铁路设计集团有限公司 Automatic track center line detection method based on vehicle-mounted mobile laser point cloud
CA3163192A1 (en) * 2019-12-17 2021-06-24 Telefonos, Lineas Y Centrales, S.A. Method for in-situ and real-time collection and processing of geometric parameters of railway lines
CN112414309A (en) * 2020-11-25 2021-02-26 北京交通大学 High-speed rail contact line height-guiding and pull-out value inspection method based on airborne laser radar

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
周靖松等: "基于三维点云的接触网几何参数检测方法", 《仪器仪表学报》 *

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114440767A (en) * 2022-01-26 2022-05-06 株洲时代电子技术有限公司 Railway comprehensive inspection method
CN114782925A (en) * 2022-06-17 2022-07-22 四川省公路规划勘察设计研究院有限公司 Vehicle-mounted LIDAR data-based highway guardrail vectorization method and device
CN114782925B (en) * 2022-06-17 2022-09-02 四川省公路规划勘察设计研究院有限公司 Highway guardrail vectorization method and device based on vehicle-mounted LIDAR data
CN115546266A (en) * 2022-11-24 2022-12-30 中国铁路设计集团有限公司 Multi-strip airborne laser point cloud registration method based on local normal correlation
CN115657067A (en) * 2022-12-26 2023-01-31 中国铁路设计集团有限公司 Multi-scale array type laser radar measuring system and method for acquiring track boundary conditions
CN116128835A (en) * 2023-02-02 2023-05-16 南方电网数字电网研究院有限公司 Point cloud analysis-based power transmission wire sag measurement method, device and equipment
CN116128835B (en) * 2023-02-02 2023-09-15 南方电网数字电网研究院有限公司 Point cloud analysis-based power transmission wire sag measurement method, device and equipment
CN115824070A (en) * 2023-02-23 2023-03-21 湖南睿图智能科技有限公司 Method and device for measuring size of steel rail piece
CN116486012A (en) * 2023-04-27 2023-07-25 中国民用航空总局第二研究所 Aircraft three-dimensional model construction method, storage medium and electronic equipment
CN116486012B (en) * 2023-04-27 2024-01-23 中国民用航空总局第二研究所 Aircraft three-dimensional model construction method, storage medium and electronic equipment
CN116862976A (en) * 2023-06-08 2023-10-10 中铁第四勘察设计院集团有限公司 Rail center line extraction method and system based on unmanned aerial vehicle laser point cloud
CN116862976B (en) * 2023-06-08 2024-04-02 中铁第四勘察设计院集团有限公司 Rail center line extraction method and system based on unmanned aerial vehicle laser point cloud

Also Published As

Publication number Publication date
CN113870123B (en) 2023-01-03
WO2023019709A1 (en) 2023-02-23

Similar Documents

Publication Publication Date Title
CN113870123B (en) Automatic detection method for contact net leading height and pulling value based on vehicle-mounted mobile laser point cloud
CN110647798B (en) Automatic track center line detection method based on vehicle-mounted mobile laser point cloud
CN111547084B (en) Data processing method based on self-moving type rail transit moving three-dimensional scanning system
CN111830528A (en) Tower characteristic point automatic identification and inclination parameter automatic measurement method based on laser point cloud
CN112414309B (en) High-speed rail contact line height-guiding and pull-out value inspection method based on airborne laser radar
CN109241069A (en) A kind of method and system that the road network based on track adaptive cluster quickly updates
CN107792115B (en) It is a kind of to automatically extract both wired rail crest level methods using three-dimensional laser point cloud
CN111666947B (en) Pantograph head offset measuring method and system based on 3D imaging
Geng et al. UAV-LiDAR-based measuring framework for height and stagger of high-speed railway contact wire
CN108564575A (en) A kind of contactless catenary's parameters detection method based on three dimensional point cloud
CN208736365U (en) Dynamic 3 D tunnel cross-section shape changing detection and analysis system
CN112036508B (en) Automatic circumferential seam identification method based on shield tunnel lining structure
CN112380902A (en) Urban rail train positioning and speed measuring method based on machine vision
CN112285738B (en) Positioning method and device for rail transit vehicle
CN114897777A (en) Full-automatic extraction method of laser point cloud of overhead line system supporting facility considering spatial relationship
CN112197743B (en) Subway tunnel contour-envelope shortest distance analysis method
CN111640155B (en) Pantograph head inclination angle measurement method and system based on 3D imaging
CN111723793B (en) Real-time rigid contact net positioning point identification method
CN113291207A (en) Dynamic measurement method of rigid subway overhead line system
CN112964195A (en) Power supply rail geometric parameter comprehensive detection method and system based on laser triangulation
CN112581521B (en) Method for extracting central line of magnetic suspension track
CN111723792B (en) Real-time positioning point identification method suitable for rigid-flexible contact network
CN111340908A (en) Track electronic map generation method
CN117848238B (en) Rail transit station actual limit measurement method based on laser point cloud
Guo et al. Frog and blade based branching direction detection in LiDAR data

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant