Research On Typical Feature Extraction Method Of True Circular Tunnel Based On 3D Laser Point Cloud Data

The subway has become the main mode of transportation for urban people.However,due to its construction below the urban surface,it has been affected by surface buildings for a long time,soil disturbance and vibration during train operation,which is easy to induce subway accidents,and thus cause serious losses to people's lives and property.Therefore,it is of great significance for the periodic health monitoring of the subway.Traditional single-point methods are now difficult to meet the high-efficiency construction and high-density maintenance requirements required for subway tunnels.The 3D laser scanning technology has been applied to the health monitoring of subway tunnels because of its automatic,high-precision and high-density information acquisition of three-dimensional coordinates on the surface of measured objects.Although 3D laser scanning can obtain a large amount of subway internal point cloud data,it also has obvious features,including complex internal information and high redundancy.If we want to realize the subway tunnel health monitoring,we need to be able to correctly extract the subway pipe wall,ensure the correctness and integrity of the data,and at the same time can effectively extract the central axis and cross-section.This paper takes the point cloud data of near section of Wenxin Subway Station of Hangzhou Metro Line 2 as an example to carry out the experiment.Systematic research on point cloud data processing for subway tunnels is achieved on MATLAB and PCL platforms.The main research as the follow:(1)The extraction of tunnel wall from point cloud data of subway.In order to accurately reflect the deformation information of the subway tunnel,it is necessary to finely extract the point cloud belonging to the pipe wall part from the redundant subway tunnel point cloud data.Based on the method of feature recognition,this paper proposes a method of extracting the subway wall by tensor voting method with random sampling consensus.Compared with the commonly used curvature extraction pipe wall method,the method can remove the noise around the pipe wall and the complete retention of the pipe wall,and provide a good data foundation for the deformation analysis of the subway tunnel.(2)The extraction and fitting of the central axis and cross-section of the subway tunnel.The central axis of the subway tunnel characterizes the posture and direction of the subway tunnel,providing a basis for the orthogonal extraction of the tunnel cross-section.Based on the noise resistance of the minimum median variance estimation algorithm,this paper applies it to the central axis extraction and fitting of metro tunnels.The section is intercepted on this basis and the Cubic b-spline method is used to perform the cross-section curve fitting.Compared with the central axis fitting method using least squares,the method is more suitable for the center point of each section and minimally affected by the outliers.(3)Case study is conducted on the point cloud data of near section of Wenxin Subway Station of Hangzhou Metro Line 2.The typical characteristics of tunnel wall,central axis and section are extracted by the method proposed in this paper,and compared with typical methods,which validate the feasibility of the proposed method.Through the experiment of the actual subway tunnel point cloud data according to the method proposed in this paper,a suitable processing method for subway point cloud data is formed.At the same time,the deformation and convergence analysis of the subway tunnel is carried out based on the extracted axis and cross-section,which effectively validates the application of 3D laser scanning technology in the health monitoring of subway tunnels.