CN118149727B - Method and system for detecting railway turnout track structure based on 3D point cloud - Google Patents
Method and system for detecting railway turnout track structure based on 3D point cloud Download PDFInfo
- Publication number
- CN118149727B CN118149727B CN202410565155.3A CN202410565155A CN118149727B CN 118149727 B CN118149727 B CN 118149727B CN 202410565155 A CN202410565155 A CN 202410565155A CN 118149727 B CN118149727 B CN 118149727B
- Authority
- CN
- China
- Prior art keywords
- data
- point cloud
- point
- rail
- measurement
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000005259 measurement Methods 0.000 claims abstract description 138
- 230000009467 reduction Effects 0.000 claims description 44
- 238000001514 detection method Methods 0.000 claims description 27
- 238000000605 extraction Methods 0.000 claims description 23
- 238000004364 calculation method Methods 0.000 claims description 22
- 238000012545 processing Methods 0.000 claims description 16
- 230000007423 decrease Effects 0.000 claims description 8
- 238000013135 deep learning Methods 0.000 claims description 8
- 238000003384 imaging method Methods 0.000 claims description 8
- 230000008859 change Effects 0.000 claims description 7
- 238000005549 size reduction Methods 0.000 claims description 7
- 238000013500 data storage Methods 0.000 claims description 6
- 230000011218 segmentation Effects 0.000 claims description 2
- 238000012423 maintenance Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 5
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 239000010410 layer Substances 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 229910052742 iron Inorganic materials 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 239000002245 particle Substances 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 238000005299 abrasion Methods 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000005452 bending Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000004880 explosion Methods 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 239000002356 single layer Substances 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/24—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B61—RAILWAYS
- B61K—AUXILIARY EQUIPMENT SPECIALLY ADAPTED FOR RAILWAYS, NOT OTHERWISE PROVIDED FOR
- B61K9/00—Railway vehicle profile gauges; Detecting or indicating overheating of components; Apparatus on locomotives or cars to indicate bad track sections; General design of track recording vehicles
- B61K9/08—Measuring installations for surveying permanent way
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Machines For Laying And Maintaining Railways (AREA)
Abstract
The invention provides a method and a system for detecting a railway turnout track structure based on 3D point clouds. Compared with the traditional manual measurement, the invention greatly reduces the measurement error caused by human factors.
Description
Technical Field
The invention relates to the technical field of rail transit detection, in particular to a method and a system for detecting a railway turnout track structure based on 3D point cloud.
Background
The high-speed railway has been developed and built in an explosive manner for nearly 20 years, and accordingly railway work also enters an explosion period of operation and maintenance requirements. The turnout is used as an important railway infrastructure, is complicated to detect and is easy to have potential safety hazards.
At present, a traditional manual measurement mode is adopted when a railway service operation and maintenance worker detects a turnout, for example, a square, a protractor and various scales are used for measuring structural parameters at the intersection of the rails, then the measurement result is manually recorded, and the manual recording measurement structure is checked with the design specification when the turnout is manufactured.
However, conventional manual measurement methods have some significant drawbacks, including, for example, accuracy, human error, measurement efficiency, and labor intensity, in particular, accuracy: manual measurement is often limited by tool precision and user skill, which may lead to larger errors in measurement results; in terms of human error: the measurement process is susceptible to subjective judgment by an operator, thereby generating errors, such as reading errors, recording errors or incorrect use of the measurement tool; in terms of measurement efficiency: manual measurements generally require more time than automated or semi-automated measurement tools, especially if a large number of repeated measurements are required; the labor intensity is as follows: manual measurements often require a great deal of manual labor, especially under harsh environmental conditions, which can lead to reduced work efficiency.
Some automatic measuring means are applied to operation and maintenance work of the rail positive line, and can be used for measuring parameters such as the rail structure size of the rail positive line. However, in the case of a switch section with a more complex structure, automated measurement is not possible.
Therefore, it is desirable to have an automated means to effectively detect switch structure dimensions, particularly reduced values, to effectively improve the quality of railroad service operations and maintenance and efficiency of the quality of the service.
Disclosure of Invention
In order to solve the defects in the prior art, the invention aims to provide a method and a system for detecting a railway turnout track structure based on 3D point cloud so as to solve the problems. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows:
in a first aspect, the present invention provides a method for detecting a railroad switch track structure based on 3D point cloud, comprising:
arranging a plurality of 3D line laser contour sensors on a measuring trolley, and simultaneously scanning and processing obtained scanning data along two tracks in a detection area, wherein the scanning data are 3D point cloud data of the tracks in the detection area;
Fusing coordinate systems of a plurality of 3D line laser contour sensors through calibration to obtain calibrated 3D point cloud data of the track, performing mileage counting based on the calibrated 3D point cloud data of the track, and performing mileage positioning on the 3D point cloud data of the track;
Carrying out gray level imaging processing on the 3D point cloud data of the calibrated track, and identifying three measurement basic points required to be subjected to structural size reduction value extraction by using a target detection model based on deep learning based on the generated gray level image, wherein the three measurement basic points are a measurement basic point based on a point rail tip, a measurement basic point based on the point rail tip and a measurement basic point based on a toe end of a guard rail;
based on the three extracted measurement basic points, respectively determining 3D point cloud data sets of measurement sections corresponding to the three measurement basic points;
determining each rail head point cloud set at the measurement section based on the 3D point cloud data sets of the measurement section corresponding to the three measurement basic points;
Determining the maximum value of the position of the stock rail along the vertical direction and the position of the vertex of the close face of the point rail according to each rail head point cloud set at the measurement section;
Calculating structure reduction values at measuring sections corresponding to the three measuring basic points, wherein the structure reduction values are equal to the maximum value of the positions of the stock rail along the vertical direction minus the vertical position of the top point of the close face of the point rail;
Comparing the structure reduction value corresponding to the three measurement basic points with a standard range corresponding to the structure reduction value, if the structure reduction value is not in the standard range corresponding to the structure reduction value, sending an alarm signal to perform dangerous position early warning, and then notifying personnel to perform manual rechecking treatment according to mileage information of the dangerous position.
Preferably, processing data acquired by scanning a plurality of 3D line laser profile sensors includes:
establishing a data receiving thread pool of the 3D line laser contour sensors, wherein each 3D line laser contour sensor corresponds to one receiving thread, and immediately storing the received data into a public data queue in the threads;
the data storage thread pool polls and reads a common data queue, which is first in first out, and stores new data into the file system immediately when the new data is found.
Preferably, the determining of each rail head point cloud set at the measurement section based on the 3D point cloud data sets of the measurement section corresponding to the three measurement basic points includes,
Dividing the whole point cloud section data at the measurement section into different data segments;
Respectively determining a minimum x value and a maximum x value of each data segment;
respectively determining the highest point of the y axis of each data segment;
and merging the data segments according to adjacent crossing rules according to the rail type characteristics to obtain each rail head point cloud set of the measurement section.
Preferably, the determining, according to each set of rail head point clouds at the measurement section, a maximum value of a position of the stock rail along the vertical direction and a position of an apex of the point rail close face includes:
Traversing from the highest point to the left or right in each determined rail head point cloud set, positioning a close joint, and displaying the change of the y-axis value of the point cloud data of the close joint from the smaller to the larger along with the decrease of the x-axis;
and continuing to traverse the section data along the x-axis based on the positioned close-fitting joint, finding out the y-axis vertex and combining the surrounding points to form the top surface of the point rail.
In a second aspect, the invention also provides a system for detecting the railway turnout track structure based on the 3D point cloud, which comprises a scanning module, a mileage positioning module, an identification module, a first extraction module, a second extraction module, a first calculation module, a second calculation module and a judgment module, wherein,
The scanning module is used for: the system comprises a plurality of 3D line laser contour sensors arranged on a measuring trolley, a detection area and a detection area, wherein the plurality of 3D line laser contour sensors are used for simultaneously scanning two tracks along the detection area and processing the obtained scanning data to obtain 3D point cloud data of the tracks in the detection area;
the mileage positioning module: the method comprises the steps of fusing coordinate systems of a plurality of 3D line laser contour sensors through calibration to obtain calibrated 3D point cloud data of a track, and performing mileage counting based on the calibrated 3D point cloud data of the track to realize mileage positioning of the 3D point cloud data of the track;
The identification module is used for carrying out gray level imaging processing on the calibrated 3D point cloud data of the track, and identifying three measurement basic points which need to be extracted with a structural size reduction value by using a target detection model based on deep learning based on the generated gray level image, wherein the three measurement basic points are a measurement basic point based on a point rail tip, a measurement basic point based on a point rail tip and a measurement basic point based on a guard rail toe end;
The first extraction module is used for respectively determining 3D point cloud data sets of measurement sections corresponding to the three measurement basic points based on the three extracted measurement basic points;
The second extraction module is used for determining each rail head point cloud set at the measurement section based on the 3D point cloud data sets of the measurement section corresponding to the three measurement basic points;
the first calculation module is used for determining the maximum value of the position of the stock rail along the vertical direction and the position of the vertex of the close face of the point rail according to each rail head point cloud set at the measurement section;
the second calculation module is used for calculating the structural reduction value at the measuring section corresponding to the three measuring basic points by using the maximum value of the position of the stock rail along the vertical direction and the vertical position of the top point of the close face of the point rail;
And the judging module is used for comparing the structure reduction values corresponding to the three measurement basic points with the standard range corresponding to the structure reduction values, sending out an alarm signal to early warn the dangerous position if the structure reduction values are not in the standard range corresponding to the structure reduction values, and informing personnel to perform manual rechecking treatment according to the point cloud coordinates of the dangerous position.
Preferably, the scanning module comprises a scanning data receiving module and a data polling module, wherein,
The scanning data receiving module: the method comprises the steps of establishing a data receiving thread pool of the 3D line laser contour sensors, wherein each 3D line laser contour sensor corresponds to one receiving thread, and immediately storing received data into a public data queue in the threads;
the data polling module: for polling a common data queue using a pool of data storage threads, the queue being first in first out, and storing new data in the file system immediately when it is found.
Preferably, the second extraction module comprises a dividing unit, a first calculation unit, a second calculation unit and a merging unit, wherein,
The dividing unit: the method comprises the steps of dividing the whole point cloud section data at a measurement section into different data segments;
the first computing unit: for determining a minimum x value and a maximum x value for each data segment, respectively;
the second calculation unit: for determining the y-axis highest point of each data segment separately;
The merging unit: and the data segments are combined according to adjacent crossing rules according to the rail type characteristics to obtain each rail head point cloud set of the measurement section.
Preferably, said first calculation module comprises a positioning unit and a point rail top surface forming unit, wherein,
The positioning unit: the method comprises the steps of traversing from the highest point to the left or right in each determined rail head point cloud set, positioning a close joint, and displaying the change of the y-axis numerical value of point cloud data of the close joint from the smaller to the larger along with the decrease of the x-axis;
the point rail top surface forming unit: for continuing to traverse the section data along the x-axis based on the located close-fitting joint, finding the y-axis vertex and combining its surrounding points to form the point rail top surface.
The invention also provides an electronic device, which comprises a memory and a processor, wherein the memory stores a program running on the processor, and the processor executes the steps of the method for detecting the railway turnout track structure based on the 3D point cloud when running the program.
In a fourth aspect, the present invention also provides a computer readable storage medium having stored thereon computer instructions that when run perform the steps of the method for detecting a railroad switch track structure based on a 3D point cloud described above.
Compared with the prior art, the method and the system for detecting the railway turnout track structure based on the 3D point cloud have the following beneficial effects:
(1) Has high precision and high resolution: the 3D point cloud technology provides high precision measurements that can accurately capture small changes in the track, including small wear, bending, or other deformations of the track.
(2) Can gather data fast: the 3D scanning device can collect a large amount of data in a short time. This speed advantage is critical to the railway system because it minimizes interference with railway operations.
(3) The safety and the reliability are better: accurate reduced value measurements are critical to ensuring the safety of railway operation, and 3D point cloud technology helps identify potential safety hazards by providing detailed track condition data.
(4) The data is easy to store and analyze: the 3D point cloud data can be stored electronically, so that the condition change of the track can be tracked and analyzed for a long time, and a basis is provided for future maintenance decisions.
(5) Support advanced data processing and modeling: the 3D point cloud data may be used to build detailed digital models of the track, supporting more advanced analysis, such as predictive maintenance using machine learning algorithms.
(6) The human error is reduced: compared with traditional manual measurement, the 3D point cloud technology greatly reduces measurement errors caused by human factors.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.
Drawings
The accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, and do not limit the invention. In the drawings:
FIG. 1 is a flow chart of a method for detecting a railroad switch track structure based on a 3D point cloud in accordance with an embodiment of the present invention;
FIG. 2 is a schematic illustration of a gray image of the tip of the point rail;
FIG. 3 is a schematic representation of a gray scale image of a tip of a heart rail;
FIG. 4 is a schematic view of a gray scale image of the toe end of the guard rail;
FIG. 5 is a schematic illustration of respective rail head point cloud gathers at a measurement cross-section according to an embodiment of the present invention;
FIG. 6 is yet another schematic illustration of various rail head point cloud sets at measurement cross-sections according to an embodiment of the present invention;
FIG. 7 is a schematic diagram of the multi-layer data phenomenon of measured cross-sectional data formed as a result of multi-camera data fitting;
FIG. 8 is a schematic diagram of a close-fitting joint;
FIG. 9 is a schematic drawing of an extraction of the top surface of a closed-loop rail head;
fig. 10 is a schematic drawing of still another closed railhead top surface extraction:
FIG. 11 is a schematic drawing of another top surface extraction of a closed rail head;
Fig. 12 is a schematic diagram of a system structure for detecting a railway switch track structure based on 3D point cloud.
Detailed Description
Embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. While the invention is susceptible of embodiment in the drawings, it is to be understood that the invention may be embodied in various forms and should not be construed as limited to the embodiments set forth herein, but rather are provided to provide a more thorough and complete understanding of the invention. It should be understood that the drawings and embodiments of the invention are for illustration purposes only and are not intended to limit the scope of the present invention.
It should be understood that the various steps recited in the method embodiments of the present invention may be performed in a different order and/or performed in parallel. Furthermore, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the invention is not limited in this respect.
The term "including" and variations thereof as used herein are intended to be open-ended, i.e., including, but not limited to. The term "based on" is based at least in part on. The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments. Related definitions of other terms will be given in the description below.
It should be noted that references to "one", "a plurality" and "a plurality" in this disclosure are intended to be illustrative rather than limiting, and those skilled in the art will appreciate that "one or more" is intended to be construed as "one or more" unless the context clearly indicates otherwise. "plurality" is understood to mean two or more.
Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
Example 1
Fig. 1 is a flowchart of a method for detecting a railway switch track structure based on a 3D point cloud according to the present invention, and the method for detecting a railway switch track structure based on a 3D point cloud according to the present invention will be described in detail with reference to fig. 1.
Referring to fig. 1, the method for detecting a railway switch track structure based on 3D point cloud of the present invention includes steps S100, S200, S300, S400, S500, S600, S700, S800.
Step S100: and arranging a plurality of 3D line laser contour sensors on the measuring trolley, and simultaneously scanning along two tracks in the detection area and processing obtained scanning data, wherein the scanning data are 3D point cloud data of the tracks in the detection area.
It should be noted that the plurality of 3D line laser profile sensors are arranged such that the total photographing range of the 3D line laser profile sensors can cover the left and right rails and simultaneously cover the complicated rail structure. Since a plurality of 3D line laser profile sensors are simultaneously scanned, a large amount of data (about 300 MB/s) is generated in time, and thus it is required to store the large amount of data in real time.
The invention adopts a double thread pool and a storage queue mode to process, and specifically comprises the steps S101 and S102:
Step S101: establishing a data receiving thread pool of the 3D line laser contour sensors, wherein each 3D line laser contour sensor corresponds to one receiving thread, and immediately storing the received data into a public data queue in the threads;
it can be appreciated that the memory operation is fast enough, and the thread returns immediately after the memory operation, so that the camera receiving unit is prevented from being blocked, and data is lost.
Step S102: the data storage thread pool polls and reads a common data queue, which is first in first out, and stores new data into the file system immediately when the new data is found.
Step S200: and fusing coordinate systems of the plurality of 3D line laser contour sensors through calibration to obtain calibrated 3D point cloud data of the track, performing mileage counting based on the calibrated 3D point cloud data of the track, and performing mileage positioning on the 3D point cloud data of the track.
It should be noted that, because each 3D line laser contour sensor has an independent coordinate system, the coordinate systems of the plurality of 3D line laser contour sensors need to be fused through calibration, so as to realize the calibration of the 3D point cloud data of the track and obtain the calibrated 3D point cloud data of the track. During calibration, the internal parameters of the 3D line laser profile sensor are calibrated first, and then external parameter calibration is needed based on a calibration table.
It will be appreciated that mileage counting may be performed by scanning the number of lines of data and the number of ties, wherein the physical distance value may be calculated using the number of lines of data, since the scanning interval of the camera is a fixed number of millimeters, and the number of ties may be used to correct the mileage value.
It is understood that mileage positioning of 3D point cloud data of a track refers to the 3D point cloud data of the track in combination with its actual mileage.
Step S300: and carrying out gray level imaging processing on the 3D point cloud data of the calibrated track, and identifying three measurement basic points which need to be subjected to structural size reduction value extraction by using a target detection model based on deep learning based on the generated gray level image, wherein the three measurement basic points are a point rail tip-based measurement basic point, a point rail tip-based measurement basic point and a guard rail toe end-based measurement basic point.
The three basic points are selected according to the requirements of the high-speed railway line equipment maintenance rule and the common-speed railway line maintenance rule (iron gauge for short).
And carrying out gray level imaging processing on the calibrated 3D point cloud data of the track, and identifying three measurement basic points which need to be extracted for the structural size reduction value by using a target detection model based on deep learning based on the generated gray level image, wherein the three measurement basic points are respectively a point rail tip-based measurement basic point, a point rail tip-based measurement basic point and a guard rail toe end-based measurement basic point. The gray image of the point tip is shown in fig. 2, wherein the red box indicated by the letter "a" indicates the point tip. The gray scale image of the tip of the heart rail is shown in fig. 3, wherein the red box indicated by the letter "B" represents the tip of the heart rail. The gray scale image of the toe end of the guard rail is shown in fig. 4, wherein the red box indicated by the letter "C" indicates the toe end of the guard rail.
It can be understood that the point cloud data only has depth data, so that the 3D point cloud data only needs to be mapped into gray values for gray imaging.
It will also be appreciated that the use of a deep learning based object detection model to identify three basic points for which a structural size reduction value extraction is required, comprises the steps of:
Firstly, in an off-line state, marking data aiming at three basic points of a point rail tip, a point rail tip and a guard rail toe end;
Then training an image semantic cutting model based on the deep neural network;
Then, during measurement, the generated gray image is input into an image semantic cutting model so as to identify three basic points in the image: when at least one base point exists, the image semantic cutting model outputs an identified base point contour.
Step S400: and respectively determining 3D point cloud data sets of measurement cross sections corresponding to the three measurement basic points based on the three extracted measurement basic points.
Based on the measurement basic points, measurement position information (relative distance difference between the measurement basic points) is acquired according to the requirement of an iron gauge, so that a measurement position is acquired, and a 3D point cloud data set of a section at the measurement position is extracted from 3D point cloud data of a rail subjected to mileage positioning.
Here, the triggering of the 3D line laser profile sensor is controlled by the odometer, so that the scanning frequency of the 3D line laser profile sensor is fixed at mileage intervals, for example, 0.138mm fixed per frame data interval. Based on the above, according to the requirement of the track, the position of the test point, that is, how many millimeters the test point is from the point rail tip (or the point rail tip), can be obtained, and converted into a specific frame number, and the frame data is extracted, so as to obtain the measurement section of the current turnout.
Step S500: and determining each rail head point cloud set at the measurement section based on the 3D point cloud data sets of the measurement section corresponding to the three measurement basic points.
It should be noted that, because the angle (ROI angle) of the region of interest (Region of Interest) acquired by the 3D line laser profile sensor is limited, each railhead data is discontinuous on the x-axis, and by using this feature, each railhead point cloud set of the measurement section can be acquired, as shown in fig. 5 and 6, where the y-axis is the track traveling direction, the x-axis is the track cross-section direction, and z is the direction perpendicular to the rail surface.
Specifically, the step of determining each rail head point cloud set at the measurement section includes steps S501, S502, S503 and S504, based on the 3D point cloud data sets of the measurement section corresponding to the three measurement basic points, wherein,
Step S501: dividing the whole point cloud section data at the measurement section into different data segments;
step S502: respectively determining a minimum x value and a maximum x value of each data segment;
step S503: respectively determining the highest point of the y axis of each data segment;
step S504: according to the rail type characteristics, the data segments are combined according to adjacent crossing rules, so that each rail head point cloud set of the measurement section is obtained, wherein the adjacent crossing rules are that continuous data points along a transverse axis are classified into the same rail head data in the measurement section, and discontinuous data points are classified into different rail head data.
Step S600: from the respective set of rail head point clouds at the measurement section, the maximum value of the position of the stock rail in the vertical direction and the position of the point contact surface vertex are determined.
It should be noted that in general, the stock rail will be higher than the point rail (or point rail), and in the case of a point rail (or point rail) in close proximity to the stock rail, the point rail surface apex can be easily determined by locating the rail head apex, but the top surface of the point rail (point rail) in close proximity is more difficult to extract. Specifically, the step of determining the top surface of the closed point rail includes steps S601, S602.
Step S601: traversing from the highest point to the left or right in each determined rail head point cloud set, positioning the close joint, and displaying the change of the y-axis numerical value of the point cloud data of the close joint from the smaller to the larger along with the decrease of the x-axis.
It should be noted that the difficulty in positioning directly to the top surface of the adjacent point rail is great because in many cross sections the features are not obvious, so here we will first position the adjacent joint, such as the circled portion of fig. 8, which is often characterized by the fact that as the x-axis decreases, the y-axis value will decrease and then increase, so we will traverse the positioning from the highest point to the left (or right).
In addition, since the cross-section data is formed by fitting a plurality of camera data, where the top surface and the peripheral position are often scanned by a plurality of cameras at the same time, that is, there are a plurality of layers of data, as shown in fig. 7, on the other hand, there are various noise points in the data, so that in the positioning process, interference in the two aspects needs to be eliminated, specifically, for a set of flying points (noise points) with a very small discrete number, the method is characterized in that: the cross section is higher in the y-axis direction, or the number of the discrete particles is extremely small in the x-axis, and the discrete particles are filtered by a set rule;
aiming at the condition of multi-layer data, the method is characterized in that: the multiple layers present in the cross-section y-axis direction may be processed by a smoothing fitting process to fit the y-axis multiple layers to single layer data.
Step S602: and continuing to traverse the section data along the x-axis based on the positioned close-fitting joint, finding out the y-axis vertex and combining the surrounding points to form the top surface of the point rail.
After the point rail is positioned at the close joint, the point rail top surface can be easily positioned, the point rail top surface can be formed by traversing the joint along the x axis, finding the y axis vertex and taking a certain amount of points around the y axis vertex. The schematic diagram of the close joint is shown in fig. 8, the schematic diagram of the top surface extraction of the close joint rail head is shown in fig. 9-11, wherein gray represents the rail section view, green represents the extracted low point position of the to-be-detected reduction value and the reduction value low rail surface (the point rail is opposite to the stock rail, the long rail is opposite to the wing rail, and the short rail is opposite to the long rail).
In addition, the units of the horizontal and vertical axes in fig. 5 to 11 are millimeters.
Step S700: and calculating structural reduction values at the measuring sections corresponding to the three measuring basic points, wherein the structural reduction values are equal to the maximum value of the positions of the stock rail along the vertical direction minus the vertical position of the top point of the contact surface of the point rail.
At the measuring sections corresponding to the three measuring basic points, the vertical position coordinates of the top points of the close surfaces of the switch rail are subtracted from the maximum value of the positions of the base rail along the vertical direction to obtain the structural reduction value of the measuring section corresponding to the measuring basic point, wherein the structural reduction value is the turnout section reduction value.
Step S800: comparing the structure reduction value corresponding to the three measurement basic points with a standard range corresponding to the structure reduction value, if the structure reduction value is not in the standard range corresponding to the structure reduction value, sending an alarm signal to perform dangerous position early warning, and then notifying personnel to perform manual rechecking treatment according to mileage information of the dangerous position.
It can be understood that, once the structural reduction value exceeds a certain range, the wheel can be inclined to incline so as to incline the vehicle body, so that the impact force of the vehicle body on the track in the fork area is enhanced, the inclination angle is increased along with the accumulation of abrasion, the running stability of the vehicle is affected slightly, and the accident is caused by the derailment of the vehicle.
When comparing the structure reduction values corresponding to the three measurement basic points with the standard range corresponding to the structure reduction values, if the structure reduction values are found not to be in the standard range corresponding to the structure reduction values, sending out an alarm signal to perform dangerous position early warning, and then notifying workers to perform manual rechecking treatment according to the point cloud coordinates of the dangerous positions: if the condition that the structure reduction value does not meet the requirement is confirmed after the manual rechecking, notifying maintenance personnel to maintain; if there is no condition that the structural degradation value does not meet the requirement, the alarm signal is eliminated and the position is marked as normal.
Example 2
As shown in fig. 12, the present embodiment provides a system for detecting a railway switch track structure based on a 3D point cloud, referring to fig. 12, which includes a scanning module 801, a mileage positioning module 802, an identification module 803, a first extraction module 804, a second extraction module 805, a first calculation module 806, a second calculation module 807, and a judgment module 808, wherein,
Scanning module 801: the system comprises a plurality of 3D line laser contour sensors arranged on a measuring trolley, a detection area and a detection area, wherein the plurality of 3D line laser contour sensors are used for simultaneously scanning two tracks along the detection area and processing the obtained scanning data to obtain 3D point cloud data of the tracks in the detection area;
Mileage positioning module 802: the method comprises the steps of fusing coordinate systems of a plurality of 3D line laser contour sensors through calibration to obtain calibrated 3D point cloud data of a track, and performing mileage counting based on the calibrated 3D point cloud data of the track to realize mileage positioning of the 3D point cloud data of the track;
The identifying module 803 is configured to perform gray level imaging processing on the calibrated 3D point cloud data of the track, identify, based on the generated gray level image, three measurement basic points that need to be extracted by using a target detection model based on deep learning, where the three measurement basic points are a measurement basic point based on a point rail tip, and a measurement basic point based on a toe end of a guard rail;
a first extraction module 804, configured to determine, based on the extracted three measurement basic points, 3D point cloud data sets of measurement cross sections corresponding to the three measurement basic points, respectively;
a second extraction module 805, configured to determine each set of rail head point clouds at the measurement section based on the 3D point cloud data sets of the measurement section corresponding to the three measurement base points;
A first calculation module 806, configured to determine, according to each set of rail head point clouds at the measurement section, a maximum value of a position of the stock rail along the vertical direction and a position of a point rail contact surface vertex;
A second calculation module 807 for calculating structural reduction values at measurement sections corresponding to the three measurement basic points using a maximum value of the position of the stock rail in the vertical direction and a vertical position of the tip point contact surface;
And the judging module 808 is configured to compare the structure reduction values corresponding to the three measurement basic points with a standard range corresponding to the structure reduction values, send out an alarm signal to perform dangerous position early warning if the structure reduction values are not in the standard range corresponding to the structure reduction values, and then notify personnel to perform manual rechecking treatment according to mileage information of the dangerous position.
Specifically, the scanning module 801 includes a scanning data receiving module 8011 and a data polling module 8012, wherein,
The scan data receiving module 8011: the method comprises the steps of establishing a data receiving thread pool of the 3D line laser contour sensors, wherein each 3D line laser contour sensor corresponds to one receiving thread, and immediately storing received data into a public data queue in the threads;
data polling module 8012: for polling a common data queue using a pool of data storage threads, the queue being first in first out, and storing new data in the file system immediately when it is found.
Specifically, the second extraction module 805 includes a dividing unit 8051, a first calculating unit 8052, a second calculating unit 8053, and a combining unit 8054, wherein,
Segmentation unit 8051: the method comprises the steps of dividing the whole point cloud section data at a measurement section into different data segments;
first calculation unit 8052: for determining a minimum x value and a maximum x value for each data segment, respectively;
the second calculation unit 8053: for determining the y-axis highest point of each data segment separately;
Merging unit 8054: and the data segments are combined according to adjacent crossing rules according to the rail type characteristics to obtain each rail head point cloud set of the measurement section.
In particular, said first calculation module 806 comprises a positioning unit 8061 and a point rail top surface forming unit 8062, wherein,
Positioning unit 8061: the method comprises the steps of traversing from the highest point to the left or right in each determined rail head point cloud set, positioning a close joint, and displaying the change of the y-axis numerical value of point cloud data of the close joint from the smaller to the larger along with the decrease of the x-axis;
Point rail top surface forming unit 8062: for continuing to traverse the section data along the x-axis based on the located close-fitting joint, finding the y-axis vertex and combining its surrounding points to form the point rail top surface.
It should be noted that, regarding the system in the above embodiment, the specific manner in which the respective modules perform the operations has been described in detail in the embodiment regarding the method, and will not be described in detail herein.
Example 3
The embodiment of the invention also provides electronic equipment, which comprises a memory and a processor, wherein the memory stores a program running on the processor, and the processor executes the steps of the method for detecting the railway turnout track structure based on the 3D point cloud when running the program. The method for detecting the railway turnout track structure based on the 3D point cloud is referred to the description of the previous part and is not repeated.
Example 4
The embodiment of the invention also provides a computer readable storage medium, wherein computer instructions are stored on the computer readable storage medium, the computer instructions execute the steps of the method for detecting the railway switch track structure based on the 3D point cloud when running, and the method for detecting the railway switch track structure based on the 3D point cloud is referred to the description of the previous parts and is not repeated.
Those of ordinary skill in the art will appreciate that: the foregoing description is only a preferred embodiment of the present invention, and the present invention is not limited thereto, but it is to be understood that modifications and equivalents of some of the technical features described in the foregoing embodiments may be made by those skilled in the art, although the present invention has been described in detail with reference to the foregoing embodiments. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A method for detecting a railroad switch track structure based on a 3D point cloud, comprising:
arranging a plurality of 3D line laser contour sensors on a measuring trolley, and simultaneously scanning and processing obtained scanning data along two tracks in a detection area, wherein the scanning data are 3D point cloud data of the tracks in the detection area;
Fusing coordinate systems of the plurality of 3D line laser contour sensors through calibration to obtain calibrated 3D point cloud data of the track, performing mileage counting based on the calibrated 3D point cloud data of the track, and performing mileage positioning on the 3D point cloud data of the track;
Carrying out gray level imaging processing on the 3D point cloud data of the calibrated track, and identifying three measurement basic points required to be subjected to structural size reduction value extraction by using a target detection model based on deep learning based on the generated gray level image, wherein the three measurement basic points are a measurement basic point based on a point rail tip, a measurement basic point based on the point rail tip and a measurement basic point based on a toe end of a guard rail;
Based on the three extracted measurement basic points, respectively determining 3D point cloud data sets of measurement sections corresponding to the three measurement basic points;
determining each rail head point cloud set at the measurement section based on the 3D point cloud data sets of the measurement section corresponding to the three measurement basic points;
determining the maximum value of the position of the stock rail along the vertical direction and the position of the point contact surface vertex according to each rail head point cloud set at the measuring section;
calculating structure reduction values at the measuring sections corresponding to the three measuring basic points, wherein the structure reduction values are equal to the maximum value of the positions of the stock rail along the vertical direction minus the vertical position of the vertex of the close face of the point rail;
Comparing the structure reduction value corresponding to the three measurement basic points with a standard range corresponding to the structure reduction value, if the structure reduction value is not in the standard range corresponding to the structure reduction value, sending an alarm signal to early warn a dangerous position, and then notifying personnel to perform manual rechecking treatment according to mileage information of the dangerous position.
2. The method for detecting railway switch track structures based on 3D point clouds as in claim 1, wherein processing data obtained by scanning said plurality of 3D line laser profile sensors comprises:
Establishing a data receiving thread pool of the 3D line laser contour sensors, wherein each 3D line laser contour sensor corresponds to one receiving thread, and immediately storing the received data into a public data queue in the threads;
the data storage thread pool polls and reads a common data queue, which is first in first out, and stores new data into the file system immediately when the new data is found.
3. The method for detecting a railroad switch track structure based on 3D point clouds as set forth in claim 1, wherein the determining each rail head point cloud set at the measurement section based on the 3D point cloud data sets of the measurement section corresponding to the three measurement base points comprises,
Dividing the whole point cloud section data at the measurement section into different data segments;
Respectively determining a minimum x value and a maximum x value of each data segment;
respectively determining the highest point of the y axis of each data segment;
and merging the data segments according to adjacent crossing rules according to the rail type characteristics to obtain each rail head point cloud set of the measurement section.
4. The method for detecting a railroad switch track structure based on 3D point clouds as set forth in claim 1, wherein the determining a maximum value of a position of the stock rail along a vertical direction and a position of a point contact surface vertex from each set of rail head point clouds at a measurement section includes:
Traversing from the highest point to the left or right in each determined rail head point cloud set, positioning a close joint, and displaying the change of the y-axis value of the point cloud data of the close joint from the smaller to the larger along with the decrease of the x-axis;
and continuing to traverse the section data along the x-axis based on the positioned close-fitting joint, finding out the y-axis vertex and combining the surrounding points to form the top surface of the point rail.
5. A system for detecting a railway turnout track structure based on 3D point cloud is characterized by comprising a scanning module, a mileage positioning module, an identification module, a first extraction module, a second extraction module, a first calculation module, a second calculation module and a judgment module, wherein,
The scanning module is used for: the system comprises a plurality of 3D line laser contour sensors arranged on a measuring trolley, a detection area and a detection area, wherein the plurality of 3D line laser contour sensors are used for simultaneously scanning two tracks along the detection area and processing the obtained scanning data to obtain 3D point cloud data of the tracks in the detection area;
the mileage positioning module: the method comprises the steps of fusing coordinate systems of a plurality of 3D line laser contour sensors through calibration to obtain calibrated 3D point cloud data of a track, and performing mileage counting based on the calibrated 3D point cloud data of the track to realize mileage positioning of the 3D point cloud data of the track;
The identification module is used for carrying out gray level imaging processing on the calibrated 3D point cloud data of the track, and identifying three measurement basic points which need to be extracted with a structural size reduction value by using a target detection model based on deep learning based on the generated gray level image, wherein the three measurement basic points are a measurement basic point based on a point rail tip, a measurement basic point based on a point rail tip and a measurement basic point based on a guard rail toe end;
The first extraction module is used for respectively determining 3D point cloud data sets of measurement sections corresponding to the three measurement basic points based on the three extracted measurement basic points;
The second extraction module is used for determining each rail head point cloud set at the measurement section based on the 3D point cloud data sets of the measurement section corresponding to the three measurement basic points;
the first calculation module is used for determining the maximum value of the position of the stock rail along the vertical direction and the position of the vertex of the close face of the point rail according to each rail head point cloud set at the measurement section;
the second calculation module is used for calculating the structural reduction value at the measuring section corresponding to the three measuring basic points by using the maximum value of the position of the stock rail along the vertical direction and the vertical position of the top point of the close face of the point rail;
And the judging module is used for comparing the structure reduction values corresponding to the three measurement basic points with the standard range corresponding to the structure reduction values, sending out an alarm signal to early warn the dangerous position if the structure reduction values are not in the standard range corresponding to the structure reduction values, and informing personnel to perform manual rechecking treatment according to the point cloud coordinates of the dangerous position.
6. The system for detecting a railroad switch track structure based on a 3D point cloud as set forth in claim 5, wherein said scanning module comprises a scanning data receiving module and a data polling module, wherein,
The scanning data receiving module: the method comprises the steps of establishing a data receiving thread pool of the 3D line laser contour sensors, wherein each 3D line laser contour sensor corresponds to one receiving thread, and immediately storing received data into a public data queue in the threads;
the data polling module: for polling a common data queue using a pool of data storage threads, the queue being first in first out, and storing new data in the file system immediately when it is found.
7. The system for detecting a railroad switch track structure based on 3D point cloud as set forth in claim 5, wherein said second extraction module comprises a segmentation unit, a first calculation unit, a second calculation unit, and a merging unit, wherein,
The dividing unit: the method comprises the steps of dividing the whole point cloud section data at a measurement section into different data segments;
the first computing unit: for determining a minimum x value and a maximum x value for each data segment, respectively;
the second calculation unit: for determining the y-axis highest point of each data segment separately;
The merging unit: and the data segments are combined according to adjacent crossing rules according to the rail type characteristics to obtain each rail head point cloud set of the measurement section.
8. The system for detecting a railroad switch rail structure based on a 3D point cloud as set forth in claim 5, wherein said first computing module comprises a positioning unit and a switch rail top surface forming unit, wherein,
The positioning unit: the method comprises the steps of traversing from the highest point to the left or right in each determined rail head point cloud set, positioning a close joint, and displaying the change of the y-axis numerical value of point cloud data of the close joint from the smaller to the larger along with the decrease of the x-axis;
the point rail top surface forming unit: for continuing to traverse the section data along the x-axis based on the located close-fitting joint, finding the y-axis vertex and combining its surrounding points to form the point rail top surface.
9. An electronic device comprising a memory and a processor, the memory having stored thereon a program running on the processor, the processor executing the steps of the method of detecting a railroad switch track structure based on a 3D point cloud of any one of claims 1-4 when the program is run.
10. A computer readable storage medium having stored thereon computer instructions which, when run, perform the steps of the method of detecting a railroad switch track structure based on a 3D point cloud as set forth in any one of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410565155.3A CN118149727B (en) | 2024-05-09 | 2024-05-09 | Method and system for detecting railway turnout track structure based on 3D point cloud |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410565155.3A CN118149727B (en) | 2024-05-09 | 2024-05-09 | Method and system for detecting railway turnout track structure based on 3D point cloud |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118149727A CN118149727A (en) | 2024-06-07 |
| CN118149727B true CN118149727B (en) | 2024-07-19 |
Family
ID=91293283
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410565155.3A Active CN118149727B (en) | 2024-05-09 | 2024-05-09 | Method and system for detecting railway turnout track structure based on 3D point cloud |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118149727B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AT5911U2 (en) * | 2002-10-29 | 2003-01-27 | Plasser Bahnbaumasch Franz | METHOD FOR CONTACT-FREE MEASUREMENT OF A CROSS-PROFILE OR DISTANCE FROM RAILS OF A TRACK |
| CN110411361A (en) * | 2019-05-15 | 2019-11-05 | 首都师范大学 | A kind of mobile tunnel laser detection data processing method |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115060188B (en) * | 2022-04-14 | 2024-07-02 | 北京交通大学 | Frog core rail molded surface processing method and system |
| DE102023103783B3 (en) * | 2023-02-16 | 2024-03-28 | Scheidt & Bachmann Gmbh | Axle counter system for monitoring a track section of a rail system |
| CN116923479A (en) * | 2023-05-19 | 2023-10-24 | 浙江银轮智能装备有限公司 | Positive line steel rail detection system and method |
| CN116934680A (en) * | 2023-05-19 | 2023-10-24 | 浙江银轮智能装备有限公司 | Method and equipment for detecting track surface block drop of turnout rail piece |
| CN117232435B (en) * | 2023-11-14 | 2024-01-30 | 北京科技大学 | Device and method for measuring abrasion value and reduction value of switch tongue |
| CN117607893A (en) * | 2023-12-04 | 2024-02-27 | 中铁第一勘察设计院集团有限公司 | Railway existing line retesting method based on unmanned aerial vehicle non-contact measurement |
-
2024
- 2024-05-09 CN CN202410565155.3A patent/CN118149727B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AT5911U2 (en) * | 2002-10-29 | 2003-01-27 | Plasser Bahnbaumasch Franz | METHOD FOR CONTACT-FREE MEASUREMENT OF A CROSS-PROFILE OR DISTANCE FROM RAILS OF A TRACK |
| CN110411361A (en) * | 2019-05-15 | 2019-11-05 | 首都师范大学 | A kind of mobile tunnel laser detection data processing method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118149727A (en) | 2024-06-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR101602376B1 (en) | A train faulty monitoring system | |
| US20120300060A1 (en) | Vision system for imaging and measuring rail deflection | |
| US20050279240A1 (en) | Enhanced method and apparatus for deducing a correct rail weight for use in rail wear analysis of worn railroad rails | |
| CN111738985B (en) | Visual detection method and system for weld joint contour | |
| ITUD980071A1 (en) | METHOD AND APPARATUS FOR DETECTING ANOMALIES IN RAILWAY AND TRAM EQUIPMENT | |
| CN110736999B (en) | Railway turnout detection method based on laser radar | |
| KR102340024B1 (en) | Track abnormality detection method applicable to various types of train tracks | |
| CN110132990A (en) | A kind of profile of steel rail detection method based on image recognition | |
| CN112960014A (en) | Rail transit operation safety online real-time monitoring and early warning management cloud platform based on artificial intelligence | |
| CN114088021B (en) | Rail straightness detection method for combined positioning of non-contact sensors | |
| CN108931206A (en) | Method for distinguishing is known for rail profile outlier detection and effective profile | |
| CN110158381B (en) | Rapid measurement method and system for track smoothness | |
| CN114312905A (en) | Switch point rail form image real-time supervision device | |
| CN113124777B (en) | Vehicle size determination method, device and system and storage medium | |
| CN118149727B (en) | Method and system for detecting railway turnout track structure based on 3D point cloud | |
| CN113091835A (en) | Network rail tunnel comprehensive detection device and method based on edge calculation and big data | |
| Rikhotso et al. | 3D rail modelling and measurement for rail profile condition assessment | |
| Chen et al. | Vision-based position deviation measurement of rail corrugation chord measuring points under bi-linear laser assistance | |
| Boronahin et al. | Optical profilometers for rail track diagnostics | |
| CN112776851B (en) | Steel rail state monitoring method and monitoring equipment | |
| CN213482904U (en) | Image analysis system and device suitable for track detection | |
| JP6389783B2 (en) | Crossover relative position management apparatus and method | |
| CN115035087A (en) | Novel railway line image detection method and system | |
| CN113963053A (en) | Method for detecting height of contact net | |
| JP7127607B2 (en) | Rail curvature estimator |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 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 |