CN118031904A - Expressway tunnel clearance measurement method and device based on vehicle-mounted laser point cloud - Google Patents

Abstract

The invention belongs to the technical field of tunnel clearance measurement, and particularly relates to a highway tunnel clearance measurement method and device based on a vehicle-mounted laser point cloud. According to the method, a vehicle-mounted three-dimensional laser scanning system is used for acquiring three-dimensional laser point cloud data in an expressway tunnel, point cloud registration is carried out on the three-dimensional laser point cloud data based on actual measurement coordinates of a layout target, filtering is carried out on the three-dimensional laser point cloud data after registration, surface layer point cloud data are extracted, a ground irregular triangular network and a lining irregular triangular network are constructed by utilizing the surface layer point cloud data, and altitude values of the lining irregular triangular network and the ground irregular triangular network are differed to obtain clearance information of the tunnel. Compared with a manual measurement method, the method has higher measurement efficiency, higher integrity of the measured tunnel clearance information and convenience for repeated measurement.

Description

Highway tunnel clearance measurement method and device based on vehicle-mounted laser point cloud

Technical Field

The invention belongs to the technical field of tunnel clearance measurement, and in particular relates to a highway tunnel clearance measurement method and device based on vehicle-mounted laser point cloud.

Background technique

Tunnel clearance refers to the space enclosed by the tunnel contour, including the cross-sectional area required for the construction limit, ventilation and other functions of the highway tunnel. The cross-sectional shape and size should strive to obtain the most economical value according to the structural design. Other cross-sectional areas included in the clearance include ventilators or ventilation ducts, lighting fixtures and other equipment, monitoring equipment and operation management equipment, cable trenches or cable trays, disaster prevention equipment, etc., as well as margins and construction tolerances.

As a particularly important part of transportation infrastructure, highway tunnels have played a very positive role in promoting the improvement of my country's national economy and rapid social development. While new highways are being built, more and more highway tunnels have been built in the past three decades. Long-term overload operation and inadequate maintenance have caused many tunnels to suffer from varying degrees of defects during operation, such as insufficient lining structure strength and cracking, pavement cracking, misalignment, bulges, insufficient pavement inverts and invert filling thickness, etc. In the process of dealing with the defects, lining I-beam steel frame reinforcement, resurfacing of the road surface and other treatment measures are usually adopted. These treatment methods will encroach on the tunnel clearance, reduce the available clearance of the lane, and pose a safety threat to vehicle driving. At the same time, due to the deformation of the tunnel, the position of the road marking may be adjusted during the process of disease treatment, and the tunnel clearance at the newly designed marking position is also a focus of attention. Therefore, the rapid, convenient and accurate extraction of the clearance of operating highway tunnels is an important part of tunnel disease treatment and evaluation of tunnel operation safety.

With the rapid development of LiDAR technology and its popular application in highway engineering, in the production process, highway tunnel clearance measurement is mainly divided into two methods. One is through manual on-the-road field measurement, using rangefinders, total stations and other measuring equipment. This method has low work efficiency, long road interruption time, and great impact on the traffic capacity of the highway network. It can only measure specific locations, and the tunnel clearance information is incomplete. The other is through three-dimensional laser scanning technology. At present, this method mostly uses station-type LiDAR scanning equipment, which has a long operation time and mainly serves the tunnel clearance convergence monitoring in the construction phase. The operation process is cumbersome and cannot meet the clearance measurement needs of repeated adjustment of tunnel markings in the operation phase.

Summary of the invention

The purpose of the present invention is to overcome the problems of low measurement efficiency, low integrity of measured tunnel clearance information and difficulty in repeated measurement in existing highway tunnel clearance measurement methods, and to provide a highway tunnel clearance measurement method and device based on vehicle-mounted laser point cloud.

In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions:

A highway tunnel clearance measurement method based on vehicle-mounted laser point cloud, the method comprises the following steps:

S1: placing a number of targets in the highway tunnel to be measured, measuring the measured coordinates of the center of each target, and using a vehicle-mounted laser scanning system to obtain multiple batches of three-dimensional laser point cloud data of the highway tunnel to be measured, wherein the three-dimensional laser point cloud data includes the point cloud data of the targets;

S2: after removing abnormal points and noise points in the three-dimensional laser point cloud data, performing point cloud registration on the three-dimensional laser point cloud data based on the measured coordinates of the centers of the targets;

S3: segment and filter the 3D laser point cloud data after point cloud registration, and calculate and generate the ground irregular triangulated network and lining irregular triangulated network;

S4: Convert the continuous line position data of the marking line position or the continuous line position data of the design line position of the highway tunnel to be tested into point data, and subtract the elevation value of the lining irregular triangulated network at the location of the point data from the elevation value of the ground irregular triangulated network to obtain the clearance value at the location of the point data.

Preferably, the point cloud registration includes performing rough point cloud registration on different batches of three-dimensional laser point cloud data, and performing fine point cloud registration on the three-dimensional laser point cloud data after the rough point cloud registration.

Preferably, the point cloud coarse registration comprises:

S21: manually marking a number of rectangular areas of point cloud data covering a single target in the current batch of three-dimensional laser point cloud data, constructing a three-dimensional statistical histogram corresponding to the rectangular area, and averaging the three-dimensional statistical histogram to obtain a reference three-dimensional statistical histogram;

S22: marking the three-dimensional laser point cloud data of the unmarked area into a plurality of rectangular areas to be matched with the same size as the rectangular area in step S21 according to a preset step size, constructing a three-dimensional statistical histogram corresponding to the rectangular area to be matched, and performing similarity matching with the reference three-dimensional statistical histogram, obtaining all rectangular areas of point cloud data covering a single target, so as to obtain the center point coordinates of all targets; wherein the center point coordinates of the target are the geometric centroid of the three-dimensional laser point cloud data within the rectangular area of point cloud data covering a single target;

S23: Correcting the coordinates of the center point of each target to be the same as the measured coordinates of the center of the target, and obtaining the three-dimensional laser point cloud data after the rough registration of the batch of point clouds.

Preferably, the method for constructing the three-dimensional statistical histogram includes: projecting the three-dimensional laser point cloud data within each rectangular area onto a plane in the direction of the normal vector of the three-dimensional laser point cloud data within the rectangular area to form a point cloud intensity image, and generating a three-dimensional statistical histogram containing grayscale and gradient information.

Preferably, the point cloud fine registration includes: calculating the mean square error of the three-dimensional laser point cloud data after the rough registration of all batches of point clouds, selecting the three-dimensional laser point cloud data of the corresponding batch when the mean square error is the smallest as the reference point cloud data, and using the ICP algorithm to register the remaining batches of three-dimensional laser point cloud data to the reference point cloud data to obtain the three-dimensional laser point cloud data after the point cloud registration.

Preferably, S3 includes: dividing the three-dimensional laser point cloud data after the point cloud registration into lower ground point cloud data and upper lining point cloud data; respectively extracting the surface point cloud data of the lower ground point cloud data and the surface point cloud data of the upper lining point cloud data, and respectively calculating and generating a ground irregular triangulated network and a lining irregular triangulated network according to the surface point cloud data.

Preferably, the surface point cloud data of the lower ground point cloud data is extracted using a progressive triangulation encryption filtering algorithm.

Preferably, the method for extracting surface point cloud data of the upper lining point cloud data comprises: extracting point cloud data below a preset elevation threshold from the upper lining point cloud data as a low-point point cloud cluster, and then using a progressive triangulation encryption filtering algorithm to extract the bottom surface point cloud data of the upper lining point cloud data, and merging the low-point point cloud cluster with the bottom surface point cloud data to obtain the surface point cloud data of the upper lining point cloud data.

Preferably, the ground irregular triangulated network and the lining irregular triangulated network are generated by using a Delaunay triangulated network growth algorithm.

A highway tunnel clearance measurement device based on vehicle-mounted laser point cloud includes at least one processor and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the above-mentioned highway tunnel clearance measurement method based on vehicle-mounted laser point cloud.

Compared with the prior art, the present invention has the following beneficial effects:

The present invention uses a three-dimensional laser scanning system to obtain three-dimensional laser point cloud data inside a highway tunnel, and performs point cloud registration on the three-dimensional laser point cloud data based on the measured coordinates of the deployed targets, and filters the registered three-dimensional laser point cloud data, extracts surface point cloud data, and uses the surface point cloud data to construct a ground irregular triangulated network and a lining irregular triangulated network, and the clearance information of the tunnel can be obtained by subtracting the elevation values of the lining irregular triangulated network and the ground irregular triangulated network. Compared with the manual measurement method, the method of the present invention can obtain the clear height information of each position in the entire tunnel area through one measurement, and the measured tunnel clearance information is more complete; in the case of line position adjustment and change, it is only necessary to re-discretize the line position and re-assign the difference according to the elevation value of the irregular triangulated network to obtain the clearance value of the changed position, which has higher measurement efficiency and is convenient for repeated measurement.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a flow chart of a highway tunnel clearance measurement method based on vehicle-mounted laser point cloud according to Embodiment 1 of the present invention;

FIG2 is a schematic diagram of a three-dimensional statistical histogram constructed in Example 1 of the present invention;

FIG3 is a three-dimensional laser point cloud data of a completely enclosed one-way highway tunnel in Example 1 of the present invention after rough registration of point clouds in the transverse cross-sectional direction;

FIG4 is a three-dimensional laser point cloud data of a completely enclosed one-way highway tunnel in Example 1 of the present invention after precise registration of point clouds in the transverse cross-sectional direction;

FIG5 is a ground irregular triangulated network constructed for a completely enclosed one-way tunnel of a highway according to Example 1 of the present invention;

FIG. 6 is an irregular triangulated network of the lining constructed for a completely enclosed one-way highway tunnel according to Example 1 of the present invention.

Detailed ways

The present invention is further described in detail below in conjunction with test examples and specific implementation methods. However, this should not be understood as the scope of the above subject matter of the present invention being limited to the following embodiments, and all technologies realized based on the content of the present invention belong to the scope of the present invention.

Example 1

The highway tunnel clearance measurement method based on vehicle-mounted laser point cloud is shown in the flowchart in Figure 1 and includes the following steps:

S1: several targets are arranged in the highway tunnel to be tested, the measured coordinates of the center of each target are measured, and a vehicle-mounted laser scanning system is used to obtain multiple batches of three-dimensional laser point cloud data of the highway tunnel to be tested, wherein the three-dimensional laser point cloud data includes point cloud data of the arranged targets.

In highway tunnels, it is difficult for the vehicle-mounted 3D laser scanning system to obtain GNSS (Global Navigation Satellite System) data, so the collected 3D laser point cloud data often relies on targets for point cloud registration; usually, the density of target layout is proportional to the length and linear complexity of the tunnel. This embodiment designs an appropriate collection speed and target layout plan based on the length, width and linearity of the highway tunnel to be measured. According to actual measurement experience, a target is usually laid out every 10 to 20 meters; and in order to obtain more dense point cloud data, the collection speed is usually inversely proportional to the width of the tunnel. After the target is laid out, the total station is used to measure the spatial three-dimensional coordinates of the center of the laid target as the measured coordinates, and the vehicle-mounted 3D laser scanning system is used to scan the tunnel to be measured to collect 3D laser point cloud data containing the point cloud data of the laid target. The model of the vehicle-mounted 3D laser scanning system used in this embodiment is AS900-HL.

For the vehicle-mounted 3D laser point cloud data collection work of highway tunnels, it is necessary to formulate differentiated collection strategies according to the different tunnel environments and operating conditions: if the tunnel to be tested is a completely closed one-way tunnel on a highway, the 3D laser point cloud data is collected back and forth in the fast and slow lanes; if it is a completely closed two-way tunnel on a highway, the 3D laser point cloud data is collected back and forth in the fast lanes in both directions; if it is a semi-closed one-way tunnel on a highway, the 3D laser point cloud data is collected back and forth in the closed lanes; if it is a semi-closed two-way tunnel on a highway, the 3D laser point cloud data is collected back and forth in the closed lanes in all directions.

A single batch of 3D laser point cloud data can be obtained by using a vehicle-mounted laser scanning system to continuously scan the entire highway tunnel to be tested in a single pass. In order to ensure the coverage integrity of the 3D laser point cloud data, at least two single-pass continuous scans are required for the highway tunnel to be tested to collect multiple batches of 3D laser point cloud data.

S2: After removing abnormal points and noise points in the three-dimensional laser point cloud data obtained in step S1, the three-dimensional laser point cloud data is aligned based on the measured coordinates of the center of each target.

The collected 3D laser point cloud data may contain abnormal points with significantly different characteristics or attributes from the surrounding points, and noise points caused by interference factors. In order to improve the quality and accuracy of the point cloud data, this embodiment adopts a neighboring point search method, that is, among all the point cloud data within the search radius, points with elevation differences greater than a certain threshold are identified as abnormal points and noise points, which are removed, and the 3D laser point cloud data after removing the noise points and abnormal points is then subjected to point cloud registration.

The point cloud registration work in this embodiment is divided into point cloud coarse registration and point cloud fine registration. First, according to the measured target center in the tunnel coordinate system measured by the total station, all batches of 3D laser point cloud data are coarsely registered. While registering the point cloud, reference point cloud data is provided for point cloud fine registration. Then, according to the reference point cloud data, the ICP algorithm is used to perform point cloud fine registration on the remaining batches of coarsely registered 3D laser point cloud data, so as to further reduce the registration error between different batches of 3D laser point cloud data and improve the quality of point cloud data.

Among them, the point cloud coarse registration is implemented in the following steps:

S21: In the current batch of three-dimensional laser point cloud data, manually mark several rectangular areas of point cloud data covering a single target, construct a three-dimensional statistical histogram corresponding to the rectangular area, and average the three-dimensional statistical histogram to obtain a benchmark three-dimensional statistical histogram.

In this embodiment, the three-dimensional laser point cloud data within each rectangular area of the point cloud data covering a single target is first projected onto a plane in the normal vector direction of the three-dimensional laser point cloud data within the rectangular area to form a point cloud intensity image, and then the image grayscale and gradient information in the point cloud intensity image are statistically analyzed to generate a three-dimensional statistical histogram. The three-dimensional statistical histograms generated by these rectangular areas are then averaged to eliminate accidental differences and obtain a reference three-dimensional statistical histogram. The three-dimensional statistical histogram containing grayscale and gradient information can better reflect the characteristics of the point cloud data within the rectangular area and improve the accuracy of similarity matching. A schematic diagram of the three-dimensional statistical histogram constructed in this embodiment is shown in FIG2 , where the grayscale on the X-axis represents the statistics of the grayscale values of the image pixels in the rectangular area marked in the point cloud intensity image, and the grayscale values are divided into 8 numerical intervals: [0,32), [32,64), [64,96), [96,128), [128,160), [160,192), [192,224), and [224-255]. The Y-axis gradient represents the statistics of the image pixel gradient direction of the circumscribed rectangle marked in the point cloud intensity image. The gradient direction is the clockwise difference between the gradient direction of a certain pixel and the direction of the image center point. The gradient direction difference is divided into 8 value intervals: [0,45), [45,90), [90,135), [135,180), [180,225), [225,270), [270,315), [315,360). The Z-axis quantity represents the number of pixels in a corresponding grayscale interval and gradient interval in the point cloud intensity image.

S22: Mark the three-dimensional laser point cloud data of the unmarked area as a number of rectangular areas to be matched with the same size as the rectangular area in step S21 according to a preset step size, construct a three-dimensional statistical histogram of the three-dimensional laser point cloud data in the rectangular area to be matched, and perform similarity matching with the reference three-dimensional statistical histogram constructed in step S21 one by one according to a preset similarity threshold. The step size set in this embodiment is 0.05 meters, and the similarity threshold is 0.9. The rectangular area above the similarity threshold is the rectangular area of the point cloud data covering a single target, and the geometric centroid of the three-dimensional laser point cloud data in these rectangular areas is the center point coordinate of the target in this embodiment.

S23: Correct the coordinates of the center point of the target obtained in S22 to be the same as the measured coordinates of the center of the target, and obtain the three-dimensional laser point cloud data after the rough registration of the batch of point clouds. In this embodiment, for a completely closed one-way tunnel of a highway, in the transverse cross-sectional direction, the three-dimensional laser point cloud data after the rough registration of the point cloud is shown in FIG3 .

After the point cloud is roughly aligned, there is still a certain error between the coordinates of the center point of the target in the 3D laser point cloud data and the measured coordinates of the target in the tunnel space coordinate system, and the errors of different targets are also different. In this embodiment, the mean error of all batches of 3D laser point cloud data after rough alignment is calculated, and the 3D laser point cloud data of the corresponding batch with the smallest mean error is selected as the reference point cloud data. The ICP algorithm is used to align the remaining batches of 3D laser point cloud data to the reference point cloud data to complete the point cloud precise alignment.

The calculation formula of mean error is as follows:

;

Among them, RMSE is the mean error of the point cloud data, n is the number of targets, _xi is the difference between the center point coordinates of the target in the 3D laser point cloud data and the measured center point coordinates of the target, It is the average value of the difference between the center point coordinates of all targets in this batch of 3D laser point cloud data and the measured center point coordinates.

The core idea of the ICP algorithm (Iterative Closest Point) is to continuously optimize the mean error between two point clouds through iteration to find the best point cloud registration result.

Specifically, in this embodiment, an initial point set _Pi is first selected from the point cloud data to be registered, and the point closest to _Pi is searched in the reference point cloud data to form a point set _Qi , so that || _Qi - _Pi || is minimized, where ||·|| represents the Euclidean distance between the two point sets; the rotation matrix R and translation matrix t required to transform the point set _Pi to the point set _Qi are calculated, and the rotation matrix R and translation matrix t are used to perform rotation and translation transformation on the point cloud data to be corrected to obtain the corrected point cloud _Pi '; the mean square error between _Pi ' and _Qi is calculated again, and if the mean square error is ≤ the set threshold, the precise registration is completed, otherwise, _Pi ' is used as the initial point set, and the above iterative calculation is performed again until the mean square error is ≤ the set threshold or the maximum number of iterations is exceeded; in this embodiment, the threshold is set to 1.5 cm, and for a completely closed one-way tunnel of a highway, the three-dimensional laser point cloud data after precise registration of the point cloud in the transverse section direction is shown in Figure 4.

S3: Segment and filter the 3D laser point cloud data after point cloud registration, and generate ground irregular triangulated network and lining irregular triangulated network.

Usually, the 3D laser point cloud data after point cloud registration is divided into lower ground point cloud data and upper lining point cloud data at 1m to 1.5m; in order to improve the accuracy of clearance information measurement, the lower ground point cloud data and the upper lining point cloud data need to be filtered to extract the surface point cloud data of the lower ground point cloud data and the surface point cloud data of the upper lining point cloud data.

Specifically, this embodiment extracts the surface point cloud data of the lower ground point cloud data and the surface point cloud data of the upper lining point cloud data based on the progressive triangulated network encryption filtering algorithm. The basic idea of the algorithm is to start from the initial ground point and gradually form the discrete point cloud data into an irregular triangulated network. This embodiment first divides the overall point cloud data into blocks, takes the point with the lowest elevation value in each block as the initial ground seed point, and constructs the initial irregular triangulated network based on this. Then, based on this initial irregular triangulated network, all the remaining point cloud data are iteratively encrypted. That is, first find the triangular face corresponding to the projection coordinates of the point to be judged on the xOy plane, calculate the slope of the triangular face, if the slope is greater than the preset terrain inclination threshold, use the mirror point of the point relative to the highest point in the triangle to determine the category of the point; if the slope of the triangular face is less than the preset terrain inclination threshold, calculate the distance from the point to be judged to the triangular face, and the angle between the line connecting the point to be judged and the vertex of the nearest triangular face and the triangular face. If the distance and the angle are both less than the preset distance threshold and angle threshold, the point is regarded as surface point cloud data, otherwise it is non-surface point cloud data. After constructing a new irregular triangulated network, repeat the above process until no new points to be judged are added or the maximum number of iterations is reached. If a point is judged as surface point cloud data, it is added to the irregular triangulated network; if it is judged as non-surface point cloud data, it is filtered out.

For the lower ground point cloud data, since there may be non-surface point cloud data such as signs, operating vehicles, and debris on the tunnel ground, it is necessary to use a progressive triangulation encryption filtering algorithm to filter and classify it; first, the lower ground point cloud data is divided into blocks, and the point with the lowest elevation value in each block is selected as the seed point to establish an initial sparse triangulation network, set the distance threshold and angle threshold, gradually encrypt the triangulation network, filter out non-surface point cloud data, and extract the surface point cloud data of the lower ground point cloud data.

For the upper lining point cloud data, since there are ventilation and lighting facilities on the top of the tunnel, if the tunnel lining has been treated for defects, there may also be I-beam steel frames and other devices. Therefore, it is necessary to first set a reasonable elevation threshold according to the conditions of the top facilities of the tunnel to be tested, and extract the point cloud data below the preset elevation threshold from the upper lining point cloud data as the low-point point cloud cluster, and then use the progressive triangulation encryption filtering algorithm to extract the bottom surface point cloud data of the upper lining point cloud data: first, the upper lining point cloud data is divided into blocks, and the point with the lowest elevation value in each block is selected as the seed point to establish the initial sparse triangulation network, set the distance threshold and angle threshold, gradually encrypt the triangulation network, filter out the non-surface point cloud data, and extract the bottom surface point cloud data of the upper lining point cloud data. Finally, the extracted low-point point cloud cluster is merged with the bottom surface point cloud data to obtain the surface point cloud data of the upper lining point cloud data.

According to the surface point cloud data of the lower ground point cloud data and the surface point cloud data of the upper lining point cloud data, the ground irregular triangulated network and the lining irregular triangulated network can be constructed respectively.

This embodiment uses the Delaunay triangulation growth algorithm to construct a ground irregular triangulation network and a lining irregular triangulation network respectively. First, any point in the extracted surface point cloud data is selected as the starting point, and the point closest to the point is connected as the initial baseline; on the right side of the initial baseline, the third point with the shortest distance to the baseline is found, and the point should satisfy that the circumscribed circle of the triangle formed by itself and the two endpoints of the baseline does not contain other points; a Delaunay triangle is generated with the two endpoints of the initial baseline and the third point found as vertices; triangles are continuously generated with the two sides of the newly generated triangulation network as new baselines until all points in the surface point cloud data are processed, and an irregular triangulation network is formed by these generated triangles. In this embodiment, for a completely closed one-way tunnel of a highway, the generated ground irregular triangulation network and lining irregular triangulation network are shown in Figures 5 and 6 respectively.

S4: Convert the continuous line position data of the marking line position or the continuous line position data of the design line position of the highway tunnel to be tested into point data, and subtract the elevation value of the lining irregular triangulated network at the location of the point data from the elevation value of the ground irregular triangulated network to obtain the clearance value at the location of the point data.

The vertices of the lining irregular triangulated network and the ground irregular triangulated network generated by the surface point cloud data have their own elevation values. Therefore, when measuring the clearance value of a certain point on the marking line position or the design line position, it is only necessary to find the triangle containing the point in the irregular triangulated network. Based on the elevation values of the vertices of the triangle corresponding to the point, the elevation value of the point in the lining irregular triangulated network and the ground irregular triangulated network is calculated using the linear interpolation method. The elevation value of the lining irregular triangulated network corresponding to the point minus the elevation value of the ground irregular triangulated network corresponding to the point is the clearance value of the location of the point. The continuous line position data of the marking line position or the continuous line position data of the design line position of the highway tunnel to be measured is converted into point data, and then the clearance values of the discrete point data on all marking line positions or design line positions are measured by the method of the present invention, so that the complete clearance information of the highway tunnel to be measured can be obtained.

Example 2

This embodiment provides a highway tunnel clearance measurement device based on vehicle-mounted laser point cloud, including at least one processor and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the highway tunnel clearance measurement method based on vehicle-mounted laser point cloud in Example 1.

The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. A highway tunnel clearance measurement method based on vehicle-mounted laser point cloud, characterized in that it includes the following steps: S1: placing a number of targets in the highway tunnel to be measured, measuring the measured coordinates of the center of each target, and using a vehicle-mounted laser scanning system to obtain multiple batches of three-dimensional laser point cloud data of the highway tunnel to be measured, wherein the three-dimensional laser point cloud data includes the point cloud data of the targets; S2: after removing abnormal points and noise points in the three-dimensional laser point cloud data, performing point cloud registration on the three-dimensional laser point cloud data based on the measured coordinates of the centers of the targets; S3: segment and filter the 3D laser point cloud data after point cloud registration, and generate ground irregular triangulated network and lining irregular triangulated network; S4: Convert the continuous line position data of the marking line position or the continuous line position data of the design line position of the highway tunnel to be tested into point data, and subtract the elevation value of the lining irregular triangulated network at the location of the point data from the elevation value of the ground irregular triangulated network to obtain the clearance value at the location of the point data. 2. The highway tunnel clearance measurement method based on vehicle-mounted laser point cloud as described in claim 1 is characterized in that the point cloud alignment includes coarse alignment of different batches of three-dimensional laser point cloud data and fine alignment of the three-dimensional laser point cloud data after the coarse alignment of the point clouds. 3. The highway tunnel clearance measurement method based on vehicle-mounted laser point cloud according to claim 2, characterized in that the point cloud coarse registration comprises: S21: manually marking a number of rectangular areas of point cloud data covering a single target in the current batch of three-dimensional laser point cloud data, constructing a three-dimensional statistical histogram corresponding to the rectangular area, and averaging the three-dimensional statistical histogram to obtain a reference three-dimensional statistical histogram; S22: marking the three-dimensional laser point cloud data of the unmarked area into a plurality of rectangular areas to be matched with the same size as the rectangular area in step S21 according to a preset step length, constructing a three-dimensional statistical histogram corresponding to the rectangular area to be matched, and performing similarity matching with the reference three-dimensional statistical histogram, obtaining all rectangular areas of point cloud data covering a single target, so as to obtain the center point coordinates of all targets; wherein the center point coordinates of the target are the geometric centroid of the three-dimensional laser point cloud data within the rectangular area of point cloud data covering a single target; S23: Correcting the coordinates of the center point of each target to be the same as the measured coordinates of the center of the target, and obtaining the three-dimensional laser point cloud data after the rough registration of the batch of point clouds. 4. The highway tunnel clearance measurement method based on vehicle-mounted laser point cloud as described in claim 3 is characterized in that the method for constructing the three-dimensional statistical histogram includes: projecting the three-dimensional laser point cloud data in each rectangular area onto a plane in the direction of the normal vector of the three-dimensional laser point cloud data in the rectangular area to form a point cloud intensity image, and generating a three-dimensional statistical histogram containing grayscale and gradient information. 5. The highway tunnel clearance measurement method based on vehicle-mounted laser point cloud as described in claim 2 is characterized in that the point cloud fine registration includes: calculating the mean square error of the three-dimensional laser point cloud data after the rough registration of all batches of point clouds, selecting the three-dimensional laser point cloud data of the corresponding batch when the mean square error is the smallest as the reference point cloud data, and using the ICP algorithm to register the remaining batches of three-dimensional laser point cloud data to the reference point cloud data to obtain the three-dimensional laser point cloud data after the point cloud registration. 6. The method for measuring clearance of a highway tunnel based on a vehicle-mounted laser point cloud as described in any one of claims 1 to 5, characterized in that S3 comprises: dividing the three-dimensional laser point cloud data after the point cloud registration into lower ground point cloud data and upper lining point cloud data; respectively extracting the surface point cloud data of the lower ground point cloud data and the surface point cloud data of the upper lining point cloud data, and respectively calculating and generating a ground irregular triangulated network and a lining irregular triangulated network according to the surface point cloud data. 7. The highway tunnel clearance measurement method based on vehicle-mounted laser point cloud as described in claim 6 is characterized in that the surface point cloud data of the lower ground point cloud data is extracted using a progressive triangulation encryption filtering algorithm. 8. The method for measuring clearance of a highway tunnel based on a vehicle-mounted laser point cloud as claimed in claim 6 is characterized in that the method for extracting surface point cloud data of the upper lining point cloud data comprises: extracting point cloud data below a preset elevation threshold from the upper lining point cloud data as a low-point point cloud cluster, then using a progressive triangulation encryption filtering algorithm to extract bottom surface point cloud data of the upper lining point cloud data, merging the low-point point cloud cluster with the bottom surface point cloud data to obtain surface point cloud data of the upper lining point cloud data. 9. The highway tunnel clearance measurement method based on vehicle-mounted laser point cloud according to claim 6, characterized in that the ground irregular triangulation network and lining irregular triangulation network are generated using a Delaunay triangulation growth algorithm. 10. A highway tunnel clearance measurement device based on vehicle-mounted laser point cloud, characterized in that it includes at least one processor and a memory communicatively connected to the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the highway tunnel clearance measurement method based on vehicle-mounted laser point cloud described in any one of claims 1 to 9.