A tunnel over-excavation and under-excavation detection method and system based on three-dimensional point cloud
Through the combination of three-dimensional point cloud technology and ICP algorithm, real-time, accuracy and visualization of tunnel ultra-under-excavation detection is achieved, solving the problems of light difference and complex operation in the existing technology, and improving construction efficiency and safety.
Patent Information
- Application Number
- CN202210025278.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-01-11
- Publication Date
- 2024-11-01
- Estimated Expiration
- 2042-01-11
AI Technical Summary
The existing tunnel ultra-under-digging detection method is inaccurate in environments with poor lighting, which affects the reconstruction accuracy, and cannot achieve online real-time evaluation, complex operation, safety hazards and material waste.
Using three-dimensional point cloud technology, through three-dimensional reconstruction of tunnels and design model point cloud registration, ICP algorithm and direct-through filtering algorithm are used to register data, convert the coordinate system and divide the grid, calculate the pole diameter difference to judge the super under-digging point , and real-time visual detection is achieved through RGB rendering.
It improves the real-time and accuracy of tunnel over-excavation inspection, simplifies the operating process, reduces the operational complexity of construction personnel, and reduces safety hazards and material waste.
Smart Images

Figure CN114359515B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of tunnel construction, and in particular to a tunnel over-break and under-break detection method and system based on three-dimensional point cloud. Background Art
[0002] During the tunnel construction process, it is difficult to avoid over-excavation and under-excavation of the section during excavation. In order to effectively control over-excavation and under-excavation, ensure the construction progress, and control the construction cost, the detection of over-excavation and under-excavation of the tunnel section is an indispensable key link in the construction process. The over-excavation and under-excavation of the tunnel is based on the designed tunnel excavation contour line as the baseline. The part of the section obtained by actual excavation outside the baseline is called over-excavation, and the part inside the baseline is called under-excavation. The consequence of over-excavation is that the mining tunnel will cause waste of materials in advance support during the later mining process. At the same time, the support effect is not good and there are safety hazards. The consequence of under-excavation is that it may cause greater trouble in transportation, pedestrians and support in the later work, and sometimes even require blasting again for correction, which also wastes materials and manpower.
[0003] The Chinese patent "CN109598714A" proposes a tunnel over-break and under-break detection method based on image 3D reconstruction and mesh surface. The real tunnel excavation contour surface is obtained by photo 3D reconstruction, and the tunnel over-break and under-break detection situation is obtained by comparing it with the meshed tunnel design excavation contour surface. However, in the tunnel environment, the lighting conditions are poor, which will seriously affect the data collection link, and then affect the reconstruction and evaluation accuracy.
[0004] The Chinese patent "CN109470206A" proposes a 3D laser scanning system for tunnel measurement. Data is collected through a 3D laser scanner, prism, total station and other data processing systems. The data preprocessing system is used to import the collected data into the software, find the position of the ball prism, and complete the point cloud positioning and orientation by rear-view orientation. After the data preprocessing is completed, the tunnel measurement evaluation is performed. This method relies on the use of a total station and cannot evaluate the over-excavation and under-excavation of the tunnel in real time online.
[0005] In terms of measuring over-break and under-break of tunnels, Zou Baoping et al. from China University of Mining and Technology proposed a method for constructing a three-dimensional visualization digital model of tunnel blasting quality based on the tunnel data acquisition module, tunnel data pre-processing module and tunnel three-dimensional visualization model construction module. The radar data was collected through the tunnel data acquisition module, and the point cloud was processed through Geomagic Studio. A three-dimensional visualization digital model of tunnel blasting quality was established, and the tunnel mining was evaluated. This method does not automatically evaluate the quality of the mining when it is completed, which brings certain operational complications and is not conducive to the operation of construction personnel.
[0006] By analyzing the above existing tunnel over-excavation and under-excavation assessment methods, it can be found that these methods have certain inaccuracies and dangers, and there is no automatic assessment of the quality of the tunnel excavation when it is completed, which brings certain operational complications and is not conducive to the operation of construction personnel. Summary of the invention
[0007] In view of the above-mentioned deficiencies in the prior art, the present invention provides a method and system for detecting over-excavation and under-excavation of a tunnel based on three-dimensional point cloud.
[0008] In order to solve the above technical problems, the technical solution adopted by the present invention is:
[0009] On the one hand, a method for detecting over-excavation and under-excavation of a tunnel based on a three-dimensional point cloud comprises the following steps:
[0010] Step 1: Use the tunnel 3D reconstruction method to obtain the 3D reconstructed tunnel point cloud data, and simultaneously obtain the model point cloud data required by the tunnel design;
[0011] Step 2: Align the 3D reconstructed lane point cloud data with the model point cloud data required by the lane design;
[0012] Furthermore, when the three-dimensionally reconstructed lane point cloud data is aligned with the model point cloud data required by the lane design, the ground and the center line of the lane in the model point cloud data required by the lane design are used as the alignment reference.
[0013] Furthermore, the three-dimensionally reconstructed tunnel point cloud data and the model point cloud data required by the tunnel design are registered and aligned using an iterative closest point registration algorithm ICP based on the ground and wall features.
[0014] Furthermore, the process of using the closest point iterative registration algorithm ICP based on the ground and wall features for registration is as follows:
[0015] S1: Set a wall height value as the extraction threshold;
[0016] S2: According to the extracted threshold, the ground and part of the wall in the three-dimensional reconstructed lane point cloud data and the model point cloud data required by the lane design are segmented using a straight-through filtering algorithm;
[0017] S3: The point cloud data required by the segmented lane design and the point cloud data of the lane after actual 3D reconstruction are registered using the closest point iterative registration algorithm ICP to obtain two point clouds with consistent postures.
[0018] Step 3: Convert the two lane point cloud data from the rectangular coordinate system to the cylindrical coordinate system;
[0019] Step 4: Divide the point cloud into grids according to the Z-axis coordinate value and polar angle value of the point cloud set after coordinate conversion, store the polar diameter value of the corresponding coordinate of each point cloud data and make a difference to obtain the coordinates of the over-excavation and under-excavation points. The specific process is as follows:
[0020] Step 4.1: Set the longitudinal segmentation value and perform longitudinal segmentation on the two registered point cloud data along the Z axis of the cylindrical coordinate system;
[0021] Step 4.2: Set the horizontal segmentation value, and perform horizontal segmentation on the two registered point cloud data based on the polar angle based on the polar point;
[0022] Step 4.3: Use the Z-axis coordinate and the polar angle as the only coordinates to store the polar diameter of the corresponding coordinates of the point cloud data in the grid, that is, each point cloud data consists of three data: polar diameter ρ, polar angle θ and z-axis length;
[0023] Step 4.4: Subtract the polar diameter of the lane design model point cloud data and the 3D reconstructed lane point cloud data in the same grid, that is, calculate ρ 极径差值 =ρ 巷道设计要求模型点云 -ρ 三维重建后的巷道点云 , where ρ 巷道设计要求模型点云 is the polar diameter value stored in the model point cloud data grid required by the roadway design, where ρ 三维重建后的巷道点云 is the polar diameter value stored in the roadway point cloud data grid after 3D reconstruction, ρ 极径差值 is the difference between the two polar diameters;
[0024] Step 4.5: When ρ 极径差值 The absolute value of is in the range [0,a], which is considered as the standard state; when ρ 极径差值 If the value of is positive and greater than a, the tunnel is judged to be under-excavated, and the coordinates of the under-excavated points are saved in the under-excavated point cloud file; when ρ 极径差值 If the value is negative and less than -a, the tunnel is judged to be in an over-excavation state, and the coordinates of the over-excavation point are saved in the over-excavation point cloud file.
[0025] Furthermore, in order to more intuitively view the over-excavation and under-excavation conditions of the tunnel, the RGB values of the under-excavation points and over-excavation points in the three-dimensionally reconstructed tunnel point cloud data are changed according to the coordinates of the over-excavation and under-excavation points obtained in step 4, respectively, to achieve the rendering of the under-excavation and over-excavation points in the three-dimensionally reconstructed tunnel point cloud data, and display them on the display.
[0026] On the other hand, a tunnel over-break and under-break detection system based on three-dimensional point cloud is provided, which applies the above-mentioned tunnel over-break and under-break detection method based on three-dimensional point cloud, and the system includes: an input module, a registration module, a detection module and a visualization module;
[0027] The input module simultaneously inputs the three-dimensional reconstruction method to obtain the three-dimensional reconstructed lane point cloud data and the model point cloud data required by the lane design;
[0028] The registration module uses the ground and the center line of the lane in the model point cloud data required by the lane design as the registration reference, and aligns the three-dimensionally reconstructed lane point cloud data with the model point cloud data required by the lane design, so that the two point clouds have a relatively consistent posture;
[0029] The detection module converts the point cloud data of the two lanes from the rectangular coordinate system to the cylindrical coordinate system; and divides the point cloud into grids according to the Z-axis coordinate value and polar angle value of the point cloud set after the coordinate conversion, stores the polar diameter values of the corresponding coordinates and makes a difference to obtain the coordinates of the over-excavation and under-excavation points;
[0030] The visualization module extracts two point clouds involved in the comparison in the grid, and respectively changes the RGB values of the under-excavated points and over-excavated points in the three-dimensionally reconstructed lane point cloud data, thereby rendering the under-excavated points and over-excavated points in the three-dimensionally reconstructed lane point cloud data, and displays them on a display.
[0031] The beneficial effect of adopting the above-mentioned technical scheme is that compared with the traditional tunnel over-excavation and under-excavation detection methods such as the total station, the method and system provided by the present invention are easy to implement and highly efficient, greatly improving the real-time performance of tunnel over-excavation and under-excavation detection, and having an intuitive visualization module, which facilitates construction personnel to detect tunnel over-excavation and under-excavation in real time. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] Figure 1 Flow chart of a method for detecting over-excavation and under-excavation of a tunnel based on three-dimensional point cloud in an embodiment of the present invention;
[0033] Figure 2 A schematic diagram of dividing point cloud data into grids in an embodiment of the present invention;
[0034] Figure 3 Schematic diagram of the structure of the tunnel over-break and under-break detection system in an embodiment of the present invention; DETAILED DESCRIPTION
[0035] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.
[0036] like Figure 1 As shown, the tunnel over-excavation and under-excavation detection method based on three-dimensional point cloud in this embodiment is described as follows.
[0037] Step 1: Use the tunnel 3D reconstruction method to obtain the 3D reconstructed tunnel point cloud data, and simultaneously obtain the model point cloud data required by the tunnel design;
[0038] Step 2: Align the 3D reconstructed lane point cloud data with the model point cloud data required by the lane design;
[0039] Furthermore, when the three-dimensionally reconstructed lane point cloud data is aligned with the model point cloud data required by the lane design, the ground and the center line of the lane in the model point cloud data required by the lane design are used as the alignment reference.
[0040] Furthermore, the three-dimensionally reconstructed tunnel point cloud data and the model point cloud data required by the tunnel design are registered and aligned using an iterative closest point registration algorithm ICP based on the ground and wall features.
[0041] Furthermore, the process of using the closest point iterative registration algorithm ICP based on the ground and wall features for registration is as follows:
[0042] S1: Set a wall height value as the extraction threshold;
[0043] In this embodiment, if the height of the segmented wall is set too high, the proportion of the ground will be reduced during registration, and the ground will not overlap, which does not meet the expected results. If the proportion of the wall is low during registration, the expected point cloud model and the actual point cloud model may be reversed during registration, which does not meet the expected results. Therefore, it is necessary to set a reasonable wall height value for segmentation. After many experiments, it was found that the actual registration effect is best when the wall height value is set to about 25% of the original wall height.
[0044] S2: According to the extracted threshold, the ground and part of the wall in the three-dimensional reconstructed lane point cloud data and the model point cloud data required by the lane design are segmented using a straight-through filtering algorithm;
[0045] S3: The point cloud data required by the segmented lane design and the point cloud data of the lane after actual 3D reconstruction are registered using the closest point iterative registration algorithm ICP to obtain two point clouds with consistent postures.
[0046] Step 3: Convert the two lane point cloud data from the rectangular coordinate system to the cylindrical coordinate system;
[0047] In this embodiment, the method of converting the rectangular coordinate system into the cylindrical coordinate system is as follows:
[0048]
[0049] z=z
[0050] Among them, x, y, and z on the right side of the equal sign are the coordinate values in the rectangular coordinate system; ρ on the left side of the equal sign is the polar diameter value of the cylindrical coordinate system, θ is the polar angle of the cylindrical coordinate system, and z is the value of the cylindrical coordinate system on the z-axis.
[0051] Step 4: Divide the point cloud into grids according to the Z-axis coordinate value and polar angle value of the point cloud set after coordinate conversion. The grid division is as follows: Figure 2 As shown, the polar diameter value of the corresponding coordinate of each point cloud data is stored and subtracted to obtain the coordinates of the over-excavation and under-excavation points. The specific process is as follows:
[0052] Step 4.1: Set the longitudinal segmentation value and perform longitudinal segmentation on the two registered point cloud data along the Z axis of the cylindrical coordinate system;
[0053] Step 4.2: Set the horizontal segmentation value, and perform horizontal segmentation on the two registered point cloud data based on the polar angle based on the polar point;
[0054] Step 4.3: Use the Z-axis coordinate and the polar angle as the only coordinates to store the polar diameter of the corresponding coordinates of the point cloud data in the grid, that is, each point cloud data consists of three data: polar diameter ρ, polar angle θ and z-axis length;
[0055] Step 4.4: Subtract the polar diameter of the lane design model point cloud data and the 3D reconstructed lane point cloud data in the same grid, that is, calculate ρ 极径差值 =ρ 巷道设计要求模型点云 -ρ 三维重建后的巷道点云 , where ρ 巷道设计要求模型点云 is the polar diameter value stored in the model point cloud data grid required by the roadway design, where ρ 三维重建后的巷道点云 is the polar diameter value stored in the roadway point cloud data grid after 3D reconstruction, ρ 极径差值 is the difference between the two polar diameters;
[0056] Step 4.5: When ρ 极径差值 The absolute value of is in the range [0,a], which is considered as the standard state; when ρ 极径差值 If the value of is positive and greater than a, the tunnel is judged to be under-excavated, and the coordinates of the under-excavated points are saved in the under-excavated point cloud file; when ρ 极径差值 If the value is negative and less than -a, the tunnel is judged to be in an over-excavation state, and the coordinates of the over-excavation point are saved in the over-excavation point cloud file.
[0057] Furthermore, in order to more intuitively view the over-excavation and under-excavation conditions of the tunnel, the RGB values of the under-excavation points and over-excavation points in the three-dimensionally reconstructed tunnel point cloud data are changed according to the coordinates of the over-excavation and under-excavation points obtained in step 4, respectively, to achieve the rendering of the under-excavation and over-excavation points in the three-dimensionally reconstructed tunnel point cloud data, and display them on the display.
[0058] On the other hand, in the embodiment of the present invention, a tunnel over-excavation and under-excavation detection system is designed for the above-mentioned tunnel over-excavation and under-excavation detection method based on three-dimensional point cloud. The system structure is as follows: Figure 3 As shown, the system includes: an input module, a registration module, a detection module and a visualization module;
[0059] The input module simultaneously inputs the three-dimensional reconstruction method to obtain the three-dimensional reconstructed lane point cloud data and the model point cloud data required by the lane design;
[0060] The registration module uses the ground and the center line of the lane in the model point cloud data required by the lane design as the registration reference, and aligns the three-dimensionally reconstructed lane point cloud data with the model point cloud data required by the lane design, so that the two point clouds have a relatively consistent posture;
[0061] The detection module converts the point cloud data of the two lanes from the rectangular coordinate system to the cylindrical coordinate system; and divides the point cloud into grids according to the Z-axis coordinate value and polar angle value of the point cloud set after the coordinate conversion, stores the polar diameter values of the corresponding coordinates and makes a difference to obtain the coordinates of the over-excavation and under-excavation points;
[0062] The visualization module extracts two point clouds involved in the comparison in the grid, and respectively changes the RGB values of the under-excavated points and over-excavated points in the three-dimensionally reconstructed lane point cloud data, thereby rendering the under-excavated points and over-excavated points in the three-dimensionally reconstructed lane point cloud data, and displays them on a display.
Claims
1. A tunnel over-break and under-break detection method based on three-dimensional point cloud, It is characterized in that The steps include: Step 1: Use the tunnel 3D reconstruction method to obtain the 3D reconstructed tunnel point cloud data, and simultaneously obtain the model point cloud data required by the tunnel design; Step 2: Align the 3D reconstructed lane point cloud data with the model point cloud data required by the lane design; Step 3: Convert the two lane point cloud data from the rectangular coordinate system to the cylindrical coordinate system; Step 4: Divide the point cloud into grids according to the Z-axis coordinate value and polar angle value of the point cloud set after coordinate conversion, store the polar diameter value of the corresponding coordinate of each point cloud data and make a difference to obtain the coordinates of the over-excavation and under-excavation points; The process of step 4 is as follows: Step 4.1: Set the longitudinal segmentation value and perform longitudinal segmentation on the two registered point cloud data along the Z axis of the cylindrical coordinate system; Step 4.2: Set the horizontal segmentation value, and perform horizontal segmentation on the two registered point cloud data based on the polar angle based on the polar point; Step 4.3: Use the Z-axis coordinate and the polar angle as the only coordinates to store the polar diameter of the corresponding coordinates of the point cloud data in the grid, that is, each point cloud data consists of three data: polar diameter ρ, polar angle θ and z-axis length; Step 4.4: Subtract the polar diameter of the lane design model point cloud data and the 3D reconstructed lane point cloud data in the same grid, that is, calculate ρ 极径差值 =ρ 巷道设计要求模型点云 -ρ 三维重建后的巷道点云 , where ρ 巷道设计要求模型点云 is the polar diameter value stored in the model point cloud data grid required by the roadway design, where ρ 三维重建后的巷道点云 is the polar diameter value stored in the roadway point cloud data grid after 3D reconstruction, ρ 极径差值 is the difference between the two polar diameters; Step 4.5: When ρ 极径差值 The absolute value of is in the range [0,a], which is considered as the standard state; when ρ 极径差值 If the value of is positive and greater than a, the tunnel is judged to be under-excavated, and the coordinates of the under-excavated points are saved in the under-excavated point cloud file; when ρ 极径差值 If the value is negative and less than -a, the tunnel is judged to be in an over-excavation state, and the coordinates of the over-excavation point are saved in the over-excavation point cloud file.
2. The tunnel over-break and under-break detection method based on three-dimensional point cloud according to claim 1, It is characterized in that When the three-dimensionally reconstructed lane point cloud data is aligned with the model point cloud data required by the lane design, the ground and the center line of the lane in the model point cloud data required by the lane design are used as the alignment reference.
3. The tunnel over-break and under-break detection method based on three-dimensional point cloud according to claim 1, It is characterized in that The three-dimensionally reconstructed lane point cloud data and the model point cloud data required by the lane design are registered and aligned using an iterative closest point registration algorithm ICP based on the ground and wall features.
4. The tunnel over-excavation and under-excavation detection method based on three-dimensional point cloud according to claim 3, It is characterized in that The process of using the closest point iterative registration algorithm ICP based on the ground and wall features for registration is as follows: S1: Set a wall height value as the extraction threshold; S2: According to the extracted threshold, the ground and part of the wall in the three-dimensional reconstructed lane point cloud data and the model point cloud data required by the lane design are segmented using a straight-through filtering algorithm; S3: The point cloud data required by the segmented lane design and the point cloud data of the lane after actual 3D reconstruction are registered using the closest point iterative registration algorithm ICP to obtain two point clouds with consistent postures.
5. The tunnel over-break and under-break detection method based on three-dimensional point cloud according to claim 1, It is characterized in that In order to more intuitively view the over-excavation and under-excavation conditions of the tunnel, the RGB values of the under-excavation points and over-excavation points in the three-dimensionally reconstructed tunnel point cloud data are changed according to the coordinates of the over-excavation and under-excavation points obtained in step 4, respectively, to achieve the rendering of the under-excavation and over-excavation points in the three-dimensionally reconstructed tunnel point cloud data, and display them on the display.
6. A tunnel over-break and under-break detection system based on three-dimensional point cloud, which uses the tunnel over-break and under-break detection method based on three-dimensional point cloud as claimed in any one of claims 1 to 5, It is characterized in that The system includes: an input module, a registration module, a detection module and a visualization module; The input module simultaneously inputs the three-dimensional reconstruction method to obtain the three-dimensional reconstructed lane point cloud data and the model point cloud data required by the lane design; The registration module uses the ground and the center line of the lane in the model point cloud data required by the lane design as the registration reference, and aligns the three-dimensionally reconstructed lane point cloud data with the model point cloud data required by the lane design, so that the two point clouds have a relatively consistent posture; The detection module converts the point cloud data of the two lanes from the rectangular coordinate system to the cylindrical coordinate system; and divides the point cloud into grids according to the Z-axis coordinate value and polar angle value of the point cloud set after the coordinate conversion, stores the polar diameter values of the corresponding coordinates and makes a difference to obtain the coordinates of the over-excavation and under-excavation points; The visualization module extracts two point clouds involved in the comparison in the grid, and respectively changes the RGB values of the under-excavated points and over-excavated points in the three-dimensionally reconstructed lane point cloud data, thereby rendering the under-excavated points and over-excavated points in the three-dimensionally reconstructed lane point cloud data, and displays them on a display.
