CN-121982051-A - Tunnel supporting structure segmentation method and system based on flatness characteristics

CN121982051ACN 121982051 ACN121982051 ACN 121982051ACN-121982051-A

Abstract

The invention provides a method and a system for segmenting a tunnel supporting structure based on flatness characteristics, which belong to the technical field of digital construction and quality detection of tunnel engineering and comprise the steps of acquiring point cloud data in a tunnel and preprocessing the data; the method comprises the steps of carrying out support structure staged feature extraction on the data of the point cloud inside the tunnel after data preprocessing, wherein primary support contour extraction and ground segmentation are carried out, namely, based on the primary support, the contour geometric feature conforming to the design line shape, the primary support contour point cloud is accurately extracted by adopting a robust model fitting method of fusion point cloud normal constraint, meanwhile, based on the plane feature of the ground, the ground point cloud is segmented and removed, and based on the smooth flatness feature of the surface of the two liners, the two-liner segmentation is carried out, namely, based on the two-stage strategy of calculating the local roughness, threshold screening and region growing of the point cloud, the two-liner point cloud is segmented from the residual point cloud after the primary support and the ground are removed, and the two-liner point cloud is segmented with high precision.

Inventors

XIE QUANYI
HU SONGYUN
LI XIAOHAN
ZHOU LIZHI
WANG QIAN
LIU JIAN
CUI LIZHUANG
ZHANG TIANTAO

Assignees

山东大学

Dates

Publication Date: 20260505
Application Date: 20260403

Claims (10)

1. The tunnel supporting structure segmentation method based on the flatness characteristics is characterized by comprising the following steps of: acquiring point cloud data in a tunnel and preprocessing the data; The method comprises the steps of carrying out support structure staged feature extraction on point cloud data in a tunnel after data preprocessing, wherein the preprocessed point cloud data is subjected to directional segmentation based on morphological features of primary support and secondary lining; The primary support contour extraction and ground segmentation are carried out, namely, a robust model fitting method of fusion point cloud normal constraint is adopted to accurately extract primary support contour point cloud based on the contour geometric characteristics of primary support conforming to design linearity; and (3) performing secondary lining segmentation, namely based on the smooth flatness characteristics of the secondary lining surface, performing high-precision segmentation on the secondary lining point cloud from the residual point cloud after primary support and ground removal by calculating a two-stage strategy of the local roughness of the point cloud, threshold screening and region growth.
2. The method for segmenting the tunnel supporting structure based on the flatness features of claim 1, further comprising the steps of logically verifying and optimizing segmentation results of primary support contour extraction and ground segmentation and secondary lining segmentation by utilizing space topology and geometric constraint relations among primary supports, secondary linings and ground, and specifically comprising the following steps: Verifying spatial correlation, namely verifying whether the secondary lining profile is reasonably contained by the primary support profile in space, and checking whether the distance between the secondary lining profile and the primary support profile accords with the designed lining thickness range; And (3) abnormal elimination and correction, namely automatically identifying and eliminating abnormal points which do not accord with the structural logic according to the verification result, or reclassifying the boundary fuzzy region, so as to ensure the correctness of the segmentation result on engineering logic.
3. The method for segmenting the tunnel supporting structure based on the flatness characteristics according to claim 1, wherein the primary support contour extraction and the ground segmentation specifically comprise the following steps: Performing cylindrical fitting, namely performing cylindrical model fitting by continuously extracting subsets from the point cloud, performing iterative optimization by using the consistency measurement of the inner points, and searching an optimal parameter solution with the maximum support samples and the minimum fitting error; Normal angle measurement, namely judging internal points based on Euclidean distance, adding joint measurement of an angle between a point cloud normal direction and a fitting model normal direction, and enhancing adaptability to a boundary detail area and a point cloud density non-uniform area; And (3) performing ground point fitting separation, namely performing local plane fitting by utilizing point clouds of an upper step area based on plane characteristics of the ground, constructing a reference plane model, realizing effective separation of ground points and elevation contours, and ensuring the integrity of ground point extraction by dynamically adjusting a distance threshold.
4. The method for segmenting the tunnel supporting structure based on the flatness characteristics according to claim 3, wherein when the reference plane model is constructed, a robust plane fitting algorithm is adopted for fitting the candidate set, so that an initial plane model is obtained, and the model is defined by plane equation parameters.
5. The method for segmenting the tunnel supporting structure based on the flatness characteristic as claimed in claim 1, wherein the two-liner segmentation specifically comprises: The roughness calculation comprises the steps of firstly adopting a voxel grid method to downsample point cloud, then calculating the local roughness of each point, defining the local roughness value as the distance between the point and a best fitting plane in the neighborhood of the point, and solving the best fitting plane parameters through covariance matrix eigenvalue decomposition; Setting a roughness threshold, screening out a point set with low roughness and uniform distribution, and extracting a point cloud with strong continuity through European clustering to finish primary segmentation; and (3) secondary segmentation of the seed points, namely searching for neighborhood points in the original high-density point cloud by taking the primary segmentation result as the seed points, and removing the duplication to realize secondary segmentation.
6. The tunnel supporting structure segmentation method based on the flatness characteristics of claim 1, wherein the specific calculation flow of calculating the local roughness of each point is that a plane equation is assumed, a covariance matrix is constructed for a target point cloud first, then feature values and corresponding feature vectors of the covariance matrix are solved, plane parameters are calculated through minimum feature values to determine a best fit plane, and finally the distance from the point to the plane is obtained to be the roughness value.
7. Tunnel supporting structure segmentation system based on roughness characteristic, characterized by includes: the data preprocessing module is configured to acquire point cloud data in the tunnel and perform data preprocessing; The method comprises the steps of carrying out support structure staged feature extraction on point cloud data in a tunnel after data preprocessing, wherein the preprocessed point cloud data is subjected to directional segmentation based on morphological features of primary support and secondary lining; The primary support contour segmentation module is configured to accurately extract primary support contour point clouds by adopting a robust model fitting method of fusion point cloud normal constraint based on the contour geometric features of primary support conforming to design linearity; The secondary lining surface segmentation module is configured to segment the secondary lining point cloud with high precision from the residual point cloud after primary support and ground are removed by calculating a two-stage strategy of point cloud local roughness, threshold screening and region growth based on the smooth flatness characteristics of the secondary lining surface.
8. A computer program product comprising a computer program, characterized in that the computer program, when executed by a processor, implements the method of any one of claims 1 to 6.
9. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the method of any of the preceding claims 1-6 when the program is executed.
10. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, performs the steps of the method of any of the preceding claims 1-6.

Description

Tunnel supporting structure segmentation method and system based on flatness characteristics Technical Field The invention belongs to the technical field of digital construction and quality detection of tunnel engineering, and particularly relates to a method and a system for dividing a tunnel supporting structure based on flatness characteristics. Background The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art. Along with the transformation of tunnel engineering construction to a digitalized and intelligent direction, the rapid and accurate monitoring and quality assessment of a tunnel construction period structure by utilizing a three-dimensional laser scanning technology become industry trends. The key structure representing different engineering stages, especially primary support and secondary lining, can be accurately separated from mass point cloud data which are acquired on site and contain complex construction interference. The primary support and the secondary lining are main permanent support structures of the tunnel, and the construction quality of the primary support and the secondary lining is directly related to the long-term safety of the tunnel. The primary support is usually applied immediately after excavation, the surface is relatively rough and may have uneven surface, the secondary lining is applied after the primary support is stable, and the surface is extremely smooth and flat after concrete pouring and template construction. This essential difference in surface morphology provides a physical basis for distinguishing the two from the point cloud. The existing point cloud segmentation method faces significant challenges when applied to the scene, wherein a deep learning-based method, such as PointNet, pointNet ++, is severely dependent on large-scale, high-quality manual annotation data for supervised training. Under the tunnel construction scene, the cost of acquiring the marking point cloud covering various geological conditions, construction stages and interference conditions is extremely high and the period is long. And when the model is used as a black box and faces complex and variable noise interference such as a construction trolley, a scaffold, a pipeline, a splashed concrete block and the like, the generalization capability is insufficient, the erroneous segmentation is easy to generate, and the stable and reliable application requirements of an engineering site are difficult to meet. In addition, methods based on conventional single geometric features, such as simple model fitting based on only the point cloud normal direction, local curvature, or using RANSAC. These methods are sensitive to noise and have poor robustness. More importantly, the complex difference between the geometric features of the outline of the primary support and the surface flatness features of the secondary lining is difficult to comprehensively describe only by relying on a single feature, so that the over-segmentation is extremely easy to divide the same structure into a plurality of parts or the under-segmentation is to combine the primary support, the secondary lining and the background noise into one part by mistake, the segmentation precision is low, and the precision requirement of engineering detection cannot be met. Disclosure of Invention In order to overcome the defects in the prior art, the invention provides a method and a system for segmenting a tunnel supporting structure based on flatness characteristics, which are used for automatically and highly precisely segmenting supporting structures such as primary supports, secondary linings and the like in tunnel construction based on the flatness or roughness characteristics of point clouds. To achieve the above object, one or more embodiments of the present invention provide the following technical solutions: In a first aspect, a method for segmenting a tunnel supporting structure based on flatness characteristics is disclosed, including: acquiring point cloud data in a tunnel and preprocessing the data; The method comprises the steps of carrying out support structure staged feature extraction on point cloud data in a tunnel after data preprocessing, wherein the preprocessed point cloud data is subjected to directional segmentation based on morphological features of primary support and secondary lining; The primary support contour extraction and ground segmentation are carried out, namely, a robust model fitting method of fusion point cloud normal constraint is adopted to accurately extract primary support contour point cloud based on the contour geometric characteristics of primary support conforming to design linearity; and (3) performing secondary lining segmentation, namely based on the smooth flatness characteristics of the secondary lining surface, performing high-precision segmentation on the secondary lining point cloud from the residual