Real-time tunnel detection method based on mobile laser scanner

Abstract

The present invention belongs to the field of tunnel detection technology, specifically a real-time tunnel detection method based on a mobile laser scanner, comprising the following steps: Step 1: laser scanner control and cross-sectional data acquisition; Step 2: scanning point coordinate conversion; Step 3: real-time limit detection; Step 4: structural deformation solution. The present invention directly completes the measurement of tunnel geometry elements and limit detection during the measurement process, and makes real-time alarms in problematic areas. Workers can review and mark on site, and directly repair the problematic areas; tunnel three-dimensional point clouds, tunnel laser intensity images, tunnel detection reports and other results can be generated in real time, greatly improving work efficiency and reducing the workload of post-processing operations; tunnel limit violations can be marked in real time in images and cross-sectional points, which is convenient for on-site workers to locate and process.

Description

Technical Field

[0001] The present invention relates to the technical field of tunnel detection, and in particular to a real-time tunnel detection method based on a mobile laser scanner. Background Art

[0002] During the operation of the tunnel, the geometric structure may be deformed by factors such as the environment, load, construction status, and train vibration, including settlement deformation, section deformation, and clearance convergence. In addition, the tunnel is restricted by the complex environment and may intrude into the normal operating limits of the track, especially at platforms, curves, switches, etc. Therefore, geometric deformation and limit detection of the tunnel is crucial to ensure the safe operation of the tunnel and the normal operation of the train.

[0003] Traditional cross-section measurement, deformation solution and limit detection based on total stations have significant disadvantages such as sparse cross-section points, low detection efficiency, low degree of automation, low safety and high labor costs. Limit measurement is also carried out by driving the limit frame on the track by the track detection vehicle. When the limit frame collides with buildings, facilities and equipment, it is considered that there is an intrusion. The detection range of this method is relatively comprehensive, but the corresponding detection frame needs to be customized for different tunnels, and the detection frame needs to be manually adjusted when the detection vehicle drives to the curved section. It has the disadvantages of high cost, low detection efficiency and poor adaptability.

[0004] Mobile laser scanners can obtain high-precision, full-section tunnel point cloud data with high detection efficiency, high degree of automation, high safety, and no need to invest a lot of manpower and material resources. Common mobile laser scanning technology is mainly based on post-processing, which requires obtaining the geometric structure parameters and intrusion conditions of the tunnel through post-processing, recording the mileage or segment information of the problem area, and returning to the site to review and handle the problem. It has poor real-time performance and a long operation cycle. Summary of the invention

[0005] 1. Technical issues to be solved

[0006] The purpose of the present invention is to solve the problems in the prior art that mobile laser scanners need to review and process problems, have poor real-time performance and a long operation cycle, and to propose a real-time tunnel detection method based on a mobile laser scanner.

[0007] 2. Technical solution

[0008] In order to achieve the above object, the present invention adopts the following technical solutions:

[0009] The tunnel real-time detection method based on mobile laser scanner includes the following steps:

[0010] Step 1: Laser scanner control and cross-sectional data collection. Use laser scanners to collect rail operating environment information. In the mobile measurement system, the laser scanner works in a two-dimensional cross-sectional mode. When the carrier moves along the tunnel, the scanning lens rotates in a direction perpendicular to the tunnel axis to form a spiral two-dimensional scanning line. When the carrier moves slowly and the laser scanning line frequency is high, the spiral scanning line data can be simplified into independent two-dimensional sections. The distance d between adjacent scanning sections is shown in formula 1.1:

[0011] d=v / f(1.1)

[0012] Each scanning point includes the two-dimensional coordinates of the point in the scanner coordinate system, laser reflection intensity, index value of the point in the cross section, availability of the point and other information. In addition, it contains relevant information of the data frame, including acquisition time, number of points and other information. After the scanner acquires a frame of data, it will respond to the data frame update event in real time and use the coordinate information and laser intensity information of the point to evaluate the availability of the point.

[0013] Step 2: Scanning point coordinate conversion. When the laser scanner is installed on a mobile measurement carrier for track measurement, the track section data obtained by the laser scanner is in the scanner coordinate system, while the limit detection is based on the track coordinate system. Therefore, the point cloud coordinates in the scan section are converted from the scanner coordinate system to the track coordinate system. The track coordinate system takes the point on the track centerline as the coordinate origin, and the X axis passes through the coordinate origin and is perpendicular to the track centerline, facing the tunnel mileage direction horizontally to the right, and the Y axis passes through the coordinate origin and is perpendicular to the track plane.

[0014] Step 3: Real-time clearance detection, including the following steps:

[0015] S3.1: Determine whether a point is an intrusion point by judging whether it is inside the bounding polygon. The horizontal ray method is selected as the intrusion point search method. A horizontal ray is drawn from the target point to the right. The point is judged whether it is inside the polygon by judging the parity of the number of intersections between the ray and the polygon. The horizontal ray method has special cases where the ray passes through the end point of the polygon or coincides with the edge of the polygon. These special cases need to be considered separately and the number of intersections needs to be corrected to avoid false detection of intrusion points.

[0016] S3.2: Flying point filtering. Due to the reflection problem of materials such as metal and glass, there will be many flying points in the rail area, platform glass screen door area, metal fence area and other areas, which are prone to false detection. Therefore, for the intrusion point, it is necessary to consider whether it is a flying point. If so, it is filtered. For dense flying points near the rail, considering that the track shape is relatively fixed, a distance threshold is set, with the midpoint of the track surface of a single track as the center of the circle. When the point appears inside the circle with the distance threshold r as the radius, it is directly filtered.

[0017] For flying points in other areas, considering that the points in the same scanning section are collected and acquired by the scanner in a certain order, even for smaller objects, the serial numbers of multiple laser points corresponding to their surfaces should be continuous, while flying points are random and often do not appear continuously. If the distance between a point and several points connected to it in sequence is close, it can be considered that the point is not a flying point, otherwise it can be regarded as a flying point. This filtering method can identify most flying points while retaining smaller objects, but there are still cases where several consecutive points are all flying points. At this time, the continuity of objects between adjacent sections is considered, that is, in two adjacent sections, objects will appear continuously at the same position, while flying points often only appear in a single section. The continuity of points in adjacent sections is judged to filter flying points. If scanning points appear at the same position in adjacent sections, they are not flying points.

[0018] S3.3: Clustering and display of limit violation points. Considering the high requirements of real-time limit detection on solution efficiency, the limit violation points are clustered based on the continuity of the points. The limit violation points are arranged according to the index value. For each limit violation point, if the index difference and distance between it and the previous limit violation point are within the threshold range, it is in the same group of clusters. Otherwise, it is considered that it is not in the same group of clusters. After clustering the limit violation point cloud, it is necessary to further describe the limit violation object, calculate the area and center coordinates of the clustered limit violation point set, and realize real-time alarm of the limit violation situation through voice to facilitate further processing by maintenance personnel.

[0019] The real-time image is generated by using the real-time acquired cross section. The laser intensity value of each point in the cross section is mapped to the range of 0-255 by stretching, and is drawn in the same column of the real-time image in sequence according to the index order of the points in the cross section. If there is an intrusion in the cross section, a red mark is added to the pixel corresponding to the intrusion area to facilitate the location of the intrusion area. The distribution of all points in the cross section is displayed in real time, and different colors are used to distinguish the intrusion points and non-intrusion points.

[0020] Step 4: Structural deformation calculation. After deformation, a circular shield tunnel is generally assumed to be an ellipse. The common indicators of circular shield tunnel deformation, including horizontal diameter and ellipticity, are calculated through the cross-sectional point set obtained in real time. First, the circle is fitted using the cross-sectional point cloud data, and the robust estimation algorithm is used to eliminate gross errors and solve the coordinates of the center of the circle. Then, the direct least squares method based on robust estimation is used to fit the ellipse. The point cloud after eliminating gross errors and the spatial distribution weighting method are used to ensure that the fitted ellipse is as close to the actual inner contour of the tunnel as possible. The algebraic equation of the general quadratic curve is formula 4.1. To ensure that the quadratic curve is an ellipse, formula 4.2 must be satisfied. The search algorithm is used to obtain the maximum diameter D passing through the center of the circle. max and minimum diameter D min , the tunnel design inner diameter D, and the ovality Ovality are calculated according to formula 4.3;

[0021] Ax 2 +Bxy+Cy 2 +Dx+Ey+F=0(4.1)

[0022] B 2 -4AC<0#(4.2)

[0023]

[0024] Preferably, in step 1, in order to achieve adaptation to scanners of different models, the control of the scanner and the reception of real-time data are abstracted into a scanner interface, which includes methods such as connecting the scanner, starting scanning, stopping scanning, disconnecting, and events of data frame update response. The data in the data frame is given in the form of a set of scanning points and is arranged in sequence along the order in which the laser is emitted by the scanning head.

[0025] Preferably, in step 1, if the distance between the point coordinates and the center of the laser scanner is less than a threshold, or the laser intensity exceeds a normal threshold range, the point is considered to be an invalid point.

[0026] Preferably, in step 2, since the scanner is usually installed perpendicular to the center line of the track, the scanner coordinate system remains parallel to the track coordinate system. It is only necessary to measure the horizontal distance and vertical distance from the origin of the scanner system to the origin of the track coordinate system, and realize the conversion of the two coordinate systems through two-dimensional coordinate translation.

[0027] Preferably, in order to ensure the accuracy and reliability of the measurement data in step 2, the device needs to be calibrated and inspected each time the scanner is installed.

[0028] Preferably, in S3.1, if the number of intersection points between the ray and the bounding polygon is an odd number, the point is inside the polygon and is judged to be intruding the limit; otherwise, it is not intruding the limit. In order to simplify the number of calculations for judgment and improve efficiency, before performing the judgment of the horizontal ray method, the coordinate range of the bounding box in the X and Y directions is first obtained. If the point is not within the range, it can be directly considered that it is not inside the polygon.

[0029] Preferably, the specific judgment method of whether the midpoint in S3.1 is inside the polygon is as follows: first check whether the target point falls on the edge of the polygon, and regard the point falling on the horizontal edge of the polygon as the internal point of the polygon, that is, the intrusion point; when the horizontal ray coincides with the edge of the polygon, only the endpoints of the edge, that is, the vertices of the polygon are regarded as intersection points; for the case where the horizontal ray intersects with the vertices of the polygon, the endpoints of each edge of the polygon are divided into upper endpoints and lower endpoints according to the size of their longitudinal coordinates. When the ray intersects the polygon at the lower endpoint, it is not regarded as an intersection; finally, judge whether the point is inside the polygon based on the parity of the corrected number of intersection points.

[0030] Preferably, if the result calculated in step 4 exceeds a set threshold, a real-time alarm of the deformation situation at that location is realized through voice to facilitate further processing by maintenance personnel.

[0031] 3. Beneficial effects

[0032] Compared with the prior art, the advantages of the present invention are:

[0033] (1) In the present invention, the two-dimensional cross-sectional mode of the laser scanner is used to work, and in combination with the carrier traveling along the tunnel, point cloud data of the entire cross-section of the tunnel is obtained in real time, the scanning points are pre-processed and the coordinates are converted, and the data in the scanner coordinate system is converted into the track coordinate system; and the relevant algorithms are used to realize the real-time detection of the deformation and limit of the tunnel geometric structure.

[0034] (2) In the present invention, the measurement of tunnel geometry and limit detection are completed directly during the measurement process, and real-time alarms are issued in problematic areas. Workers can review and mark on site and directly repair the problematic areas. In addition, tunnel three-dimensional point clouds, tunnel laser intensity images, tunnel detection reports and other results can be generated in real time, greatly improving work efficiency and reducing the workload of post-processing operations. Tunnel limit violations can be marked in real time in images and cross-section points, making it convenient for on-site workers to locate and handle them. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 Diagram of the real-time tunnel monitoring system based on mobile laser scanning proposed by the present invention;

[0036] Figure 2 Schematic diagram of the two-dimensional spiral proposed by the present invention;

[0037] Figure 3 Positional relationship diagram of the scanner coordinate system and the track coordinate system proposed by the present invention;

[0038] Figure 4 Principle diagram of the horizontal ray method proposed by the present invention;

[0039] Figure 5 The present invention proposes a flying point filtering diagram near the track;

[0040] Figure 6 The flying point filtering principle diagram based on continuous points proposed by the present invention;

[0041] Figure 7 The principle diagram of the flying point filtering based on continuous sections proposed by the present invention;

[0042] Figure 8 A diagram of the real-time detection interface proposed by the present invention. DETAILED DESCRIPTION

[0043] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0044] Embodiment 1:

[0045] Reference Figure 1-8 , a tunnel real-time detection method based on a mobile laser scanner includes the following steps:

[0046] Step 1: Laser scanner control and cross-sectional data collection. Use laser scanners to collect rail operating environment information. In the mobile measurement system, the laser scanner works in a two-dimensional cross-sectional mode. When the carrier moves along the tunnel, the scanning lens rotates in a direction perpendicular to the tunnel axis, forming a spiral two-dimensional scanning line. When the carrier moves slowly and the laser scanning line frequency is high, the spiral scanning line data can be simplified into independent two-dimensional sections. The distance d between adjacent scanning sections can be seen in the formula:

[0047] d=v / f(1.1)

[0048] In order to adapt to different models of scanners, the control of the scanner and the reception of real-time data are abstracted into a scanner interface. The interface includes methods such as scanner connection, start scanning, stop scanning, disconnection, and data frame update response events. The data in the data frame is given in the form of a set of scanning points and is arranged in the order in which the scanner emits lasers. Each scanning point includes the two-dimensional coordinates of the point in the scanner coordinate system, the laser reflection intensity, the index value of the point in the cross section, the availability of the point, and other information. In addition, it contains relevant information of the data frame, including acquisition time, number of points, etc. After the scanner acquires a frame of data, it will respond to the data frame update event in real time. The coordinate information and laser intensity information of the point are used to evaluate the availability of the point. If the distance between the point coordinate and the center of the laser scanner is less than the threshold, or the laser intensity exceeds the normal threshold range, the point is considered to be an invalid point, and the availability attribute of the point is modified to no. When implementing scanner adaptation, it is only necessary to carry out secondary development for different scanner models and inherit the interface. Currently, the adaptation work of scanners such as Z+F 9012, Z+F 9020, Sick LMS111, Sick LMS511, Sick LMS4000 series, and Leica P series has been completed, and it has good applicability;

[0049] Step 2: Scanning point coordinate conversion. When the laser scanner is installed on a mobile measuring carrier for track measurement, the track section data obtained by the laser scanner is in the scanner coordinate system, while the limit detection is based on the track coordinate system. To this end, the point cloud coordinates in the scanning section are converted from the scanner coordinate system to the track coordinate system. The track coordinate system uses the point on the track centerline as the coordinate origin. The X-axis passes through the coordinate origin and is perpendicular to the track centerline. It faces the tunnel's long mileage direction and is horizontal to the right. The Y-axis passes through the coordinate origin and is perpendicular to the track plane. Since the scanner is usually installed perpendicular to the track centerline, the scanner coordinate system remains parallel to the track coordinate system. It is only necessary to measure the horizontal and vertical distances from the scanner system origin to the track coordinate system origin, and realize the conversion of the two coordinate systems through two-dimensional coordinate translation. To ensure the accuracy and reliability of the measurement data, the equipment needs to be calibrated and inspected each time the scanner is installed;

[0050] Step 3: Real-time clearance detection, including the following steps:

[0051] S3.1: Determine whether the point is an intrusion point by judging whether it is inside the bounding polygon. The horizontal ray method is selected as the intrusion point search method. Draw a horizontal ray from the target point to the right, and judge whether the point is inside the polygon by judging the parity of the number of intersections between the ray and the polygon. If the number of intersections between the ray and the bounding polygon is odd, the point is inside the polygon and is judged to be an intrusion; otherwise, it is not an intrusion. In order to simplify the number of calculations for judgment and improve efficiency, before the horizontal ray method is used for judgment, the coordinate range of the bounding box in the X and Y directions is first obtained. If the point is not within the range, it can be directly considered to be outside the polygon.

[0052] The horizontal ray method has special cases where the ray passes through the endpoints of the polygon or coincides with the edge of the polygon, which need to be considered separately. The number of intersection points is corrected to avoid misjudgment of limit-invading points. The specific judgment method is as follows: first, check whether the target point falls on the edge of the polygon, and regard the points that fall on the horizontal edge of the polygon as internal points of the polygon, that is, limit-invading points; when the horizontal ray coincides with the edge of the polygon, only the endpoints of the edge (vertices of the polygon) are regarded as intersection points; for the case where the horizontal ray intersects with the vertex of the polygon, the endpoints of each edge of the polygon are divided into upper endpoints and lower endpoints according to the size of their ordinates. When the ray intersects the polygon at the lower endpoint, it is not considered to intersect; finally, judge whether the point is inside the polygon based on the parity of the corrected number of intersection points.

[0053] S3.2: Flying point filtering. Due to the reflection problem of metal, glass and other materials, there will be many flying points in the rail area, platform glass shielding door area, metal fence area and other areas, which are prone to false detection. Therefore, for the intrusion point, it is necessary to consider whether it is a flying point. If so, it is filtered.

[0054] For dense flying points near the rails, considering that the track shape is relatively fixed, a distance threshold is set, with the midpoint of the track surface of a single track as the center of the circle. When a point appears inside the circle with the distance threshold r as the radius, it is directly filtered.

[0055] For flying points in other areas, considering that the points in the same scanning section are collected and acquired by the scanner in a certain order, even for smaller objects, the serial numbers of multiple laser points corresponding to their surfaces should be continuous, while flying points are random and often do not appear continuously. If the distance between a point and several points connected to it in sequence is close, it can be considered that the point is not a flying point, otherwise it can be regarded as a flying point. This filtering method can identify most flying points while retaining smaller objects, but there are still cases where several consecutive points are all flying points. At this time, the continuity of objects between adjacent sections is considered, that is, in two adjacent sections, objects will appear continuously at the same position, while flying points often only appear in a single section. The continuity of points in adjacent sections is judged to filter flying points. If scanning points appear at the same position in adjacent sections, they are not flying points.

[0056] This method only needs to determine the proximity of several adjacent points to correctly filter out most of the flying points, ensuring the efficiency and reliability of detection and achieving real-time detection of the limit. For intruders whose length is less than the interval between adjacent scanning sections, this method may result in missed detection. This problem can be improved by reducing the running speed of the carrier or increasing the scanning frequency of the scanner.

[0057] S3.3: Clustering and display of limit violation points. Considering the high requirements of real-time limit detection on solution efficiency, the clustering of limit violation points is based on the continuity of points. Arrange the limit violation points according to the index value. For each limit violation point, if the index difference and distance between it and the previous limit violation point are within the threshold range, it is in the same group of clusters, otherwise it is considered that it is not in the same group of clusters. After clustering the limit violation point cloud, it is necessary to further describe the limit violation object. Calculate the area and center coordinates of the clustered limit violation point set, and realize real-time alarm of the limit violation situation through voice to facilitate further processing by maintenance personnel.

[0058] The real-time image is generated by using the real-time acquired cross section. The laser intensity value of each point in the cross section is mapped to the range of 0-255 by stretching, and is drawn in the same column of the real-time image in the order of the index of the points in the cross section. If there is an intrusion in the cross section, a red mark is added to the pixel corresponding to the intrusion area to facilitate the location of the intrusion area. The distribution of all points in the cross section is displayed in real time, and different colors are used to distinguish the intrusion points and non-intrusion points;

[0059] Step 4: Structural deformation calculation. After deformation, a circular shield tunnel is generally assumed to be an ellipse. The common indicators of circular shield tunnel deformation, including horizontal diameter and ellipticity, are calculated through the cross-sectional point set obtained in real time. First, the circle is fitted using the cross-sectional point cloud data, and the robust estimation algorithm is used to eliminate gross errors and solve the coordinates of the center of the circle; then the direct least squares method based on robust estimation is used to fit the ellipse. The point cloud after eliminating gross errors and the spatial distribution weighting method are used to ensure that the fitted ellipse fits the actual inner contour of the tunnel as closely as possible. The algebraic equation of a general quadratic curve is Equation 4.1. To ensure that the quadratic curve is an ellipse, Equation 4.2 must be satisfied. A search algorithm is used to obtain the maximum diameter D passing through the center of the circle. max and minimum diameter D min , the tunnel design inner diameter D, and the ovality Ovality are calculated according to formula 4.3. If the calculated result exceeds the set threshold, a real-time alarm of the deformation situation at that location will be realized through voice to facilitate further processing by maintenance personnel.

[0060] Ax 2 +Bxy+Cy 2 +Dx+Ey+F=0(4.1)

[0061] B 2 -4AC<0#(4.2)

[0062]

[0063] In the present invention, the two-dimensional cross-sectional mode of the laser scanner is utilized, and in combination with the carrier traveling along the tunnel, point cloud data of the entire cross-section of the tunnel is acquired in real time, the scanning points are preprocessed and the coordinates are transformed, and the data in the scanner coordinate system is converted into the track coordinate system; and the real-time detection of the deformation and limit of the tunnel geometric structure is realized through the relevant algorithms.

[0064] In the present invention, the measurement of tunnel geometry and limit detection are completed directly during the measurement process, and real-time alarms are issued in problematic areas. Workers can review and mark on site and directly perform maintenance on problematic areas. In addition, tunnel three-dimensional point clouds, tunnel laser intensity images, tunnel detection reports and other results can be generated in real time, greatly improving work efficiency and reducing the workload of post-processing operations. Tunnel limit violations can be marked in real time in images and section points, making it convenient for on-site workers to locate and handle them.

[0065] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A real-time tunnel detection method based on a mobile laser scanner, characterized in that: The following steps are involved: Step 1: Laser scanner control and cross-sectional data collection. Use laser scanners to collect rail operating environment information. In the mobile measurement system, the laser scanner works in a two-dimensional cross-sectional mode. When the carrier moves along the tunnel, the scanning lens rotates in a direction perpendicular to the tunnel axis to form a spiral two-dimensional scanning line. When the carrier moves at a slow speed v and the laser scanning line frequency f is high, the spiral scanning line data can be simplified into independent two-dimensional sections. The distance between adjacent scanning sections See formula 1.1: Each scanning point includes the two-dimensional coordinates of the point in the scanner coordinate system, the laser reflection intensity, the index value of the point in the cross section, and the availability information of the point. In addition, it contains relevant information of the data frame, including the acquisition time and the number of points. After the scanner acquires a frame of data, it will respond to the data frame update event in real time and use the coordinate information and laser intensity information of the point to evaluate the availability of the point. Step 2: Scanning point coordinate conversion. When the laser scanner is installed on a mobile measurement carrier for track measurement, the track section data obtained by the laser scanner is in the scanner coordinate system, while the limit detection is based on the track coordinate system. Therefore, the point cloud coordinates in the scan section are converted from the scanner coordinate system to the track coordinate system. The track coordinate system takes the point on the track centerline as the coordinate origin, and the X axis passes through the coordinate origin and is perpendicular to the track centerline, facing the tunnel mileage direction horizontally to the right, and the Y axis passes through the coordinate origin and is perpendicular to the track plane. Step 3: Real-time clearance detection, including the following steps: S3.1: Determine whether a point is an intrusion point by judging whether it is inside the bounding polygon. The horizontal ray method is selected as the intrusion point search method. A horizontal ray is drawn from the target point to the right. The point is judged whether it is inside the polygon by judging the parity of the number of intersections between the ray and the polygon. The horizontal ray method has special cases where the ray passes through the end point of the polygon or coincides with the edge of the polygon. These special cases need to be considered separately and the number of intersections needs to be corrected to avoid false detection of intrusion points. S3.2: Flying point filtering. Due to the reflection problem of metal and glass materials, there will be many flying points in the rail area, platform glass screen door area, and metal fence area, which are prone to false detection. Therefore, for the intrusion point, it is necessary to consider whether it is a flying point. If so, it is filtered. For dense flying points near the rail, considering that the track shape is relatively fixed, a distance threshold is set, with the midpoint of the track surface of a single track as the center of the circle. When the point appears inside the circle with the distance threshold r as the radius, it is directly filtered. For flying points in other areas, considering that the points in the same scanning section are collected and acquired by the scanner in a certain order, even for smaller objects, the serial numbers of multiple laser points corresponding to their surfaces should be continuous, while flying points are random and often do not appear continuously. If the distance between a point and several points connected to it in sequence is close, it can be considered that the point is not a flying point, otherwise it is considered a flying point. This filtering method identifies most flying points while retaining smaller objects, but there are still cases where several consecutive points are all flying points. At this time, the continuity of objects between adjacent sections is considered, that is, in two adjacent sections, objects will appear continuously at the same position, while flying points often appear only in a single section. The continuity of points in adjacent sections is judged to filter flying points. If scanning points appear at the same position in adjacent sections, they are not flying points. S3.3: Clustering and display of limit violation points. Considering the high requirements of real-time limit detection on solution efficiency, the limit violation points are clustered based on the continuity of the points. The limit violation points are arranged according to the index value. For each limit violation point, if the index difference and distance between it and the previous limit violation point are within the threshold range, it is in the same group of clusters. Otherwise, it is considered that it is not in the same group of clusters. After clustering the limit violation point cloud, it is necessary to further describe the limit violation object, calculate the area and center coordinates of the clustered limit violation point set, and realize real-time alarm of the limit violation situation through voice to facilitate further processing by maintenance personnel. The real-time image is generated by using the real-time acquired cross section. The laser intensity value of each point in the cross section is mapped to the range of 0-255 by stretching, and is drawn in the same column of the real-time image in sequence according to the index order of the points in the cross section. If there is an intrusion in the cross section, a red mark is added to the pixel corresponding to the intrusion area to facilitate the location of the intrusion area. The distribution of all points in the cross section is displayed in real time, and different colors are used to distinguish the intrusion points and non-intrusion points. Step 4: Structural deformation calculation. The circular shield tunnel is assumed to be an ellipse after deformation. The common indicators of circular shield tunnel deformation, including horizontal diameter and ellipticity, are calculated through the cross-sectional point set obtained in real time. First, the circle is fitted using the cross-sectional point cloud data, and the robust estimation algorithm is used to eliminate gross errors and solve the coordinates of the center of the circle. Then, the direct least squares method based on robust estimation is used to fit the ellipse. The point cloud after eliminating gross errors and the spatial distribution weighting method are used to ensure that the fitted ellipse fits the actual inner contour of the tunnel as much as possible. The algebraic equation of the quadratic curve is Formula 4.

1. To ensure that the quadratic curve is an ellipse, Formula 4.2 must be satisfied. The search algorithm is used to obtain the maximum diameter through the center of the circle. and minimum diameter , tunnel design inner diameter 1. Ellipticity Calculate according to formula 4.3; 。 2. The real-time tunnel detection method based on a mobile laser scanner according to claim 1 is characterized in that: In step 1, in order to achieve adaptation to scanners of different models, the control of the scanner and the reception of real-time data are abstracted into a scanner interface. The interface includes events of scanner connection, scanning start, scanning stop, disconnection method and data frame update response. The data in the data frame is given in the form of a set of scanning points and is arranged in sequence along the order in which the laser is emitted by the scanning head.

3. The real-time tunnel detection method based on mobile laser scanner according to claim 1 is characterized in that: In step 1, if the distance between the point coordinates and the center of the laser scanner is less than a threshold, or the laser intensity exceeds a normal threshold range, the point is considered to be an invalid point.

4. The real-time tunnel detection method based on a mobile laser scanner according to claim 1 is characterized in that: In step 2, since the scanner is installed perpendicular to the center line of the track, the scanner coordinate system remains parallel to the track coordinate system. It is only necessary to measure the horizontal and vertical distances from the origin of the scanner system to the origin of the track coordinate system, and realize the conversion of the two coordinate systems through two-dimensional coordinate translation.

5. The real-time tunnel detection method based on mobile laser scanner according to claim 1 is characterized in that: In order to ensure the accuracy and reliability of the measurement data in step 2, the device needs to be calibrated and inspected each time the scanner is installed.

6. The real-time tunnel detection method based on mobile laser scanner according to claim 1 is characterized in that: In S3.1, if the number of intersections between the ray and the bounding polygon is an odd number, the point is inside the polygon and is judged to be intrusive; otherwise, it is not intrusive. In order to simplify the number of calculations for judgment and improve efficiency, before performing the judgment of the horizontal ray method, the coordinate range of the bounding box in the X and Y directions is first obtained. If the point is not within the range, it can be directly considered that it is not inside the polygon.

7. The real-time tunnel detection method based on a mobile laser scanner according to claim 6 is characterized in that: The specific method for determining whether the midpoint in S3.1 is inside the polygon is as follows: first, check whether the target point falls on the edge of the polygon, and regard the point falling on the horizontal edge of the polygon as the internal point of the polygon, that is, the intrusion point; when the horizontal ray coincides with the edge of the polygon, only the endpoint of the edge, that is, the vertex of the polygon, is regarded as the intersection point; in the case where the horizontal ray intersects with the vertex of the polygon, the endpoint of each edge of the polygon is divided into an upper endpoint and a lower endpoint according to the size of its ordinate, and when the ray intersects the polygon at the lower endpoint, it is not regarded as an intersection; Finally, determine whether the point is inside the polygon based on the parity of the corrected number of intersection points.

8. The real-time tunnel detection method based on mobile laser scanner according to claim 1 is characterized in that: If the result calculated in step 4 exceeds the set threshold, a real-time alarm of the deformation situation is realized through voice to facilitate further processing by maintenance personnel.

Images