基于椭圆拟合的隧道点云数据去噪方法

针对地铁隧道这一狭长密闭环境,激光雷达相较于传统隧道形变检测方法具有显著优势。由于受环境以及隧道中安装的大量设备的影响,数据中会包含大量噪声点,无法得到隧道完整内表壁数据,从而影响到数据在形变检测及三维建模等方面的应用。结合最小二乘法和拉格朗日乘数法对隧道点云模型进行椭圆拟合,并将其应用于圆形隧道点云滤波,利用各区域数据对隧道形变进行检测。对滤波结果的精度评定及其有效性分析表明,此算法针对隧道点云数据的滤波具有良好效果及稳定性。

Denoising method of tunnel point cloud data based on ellipse fitting

In view of the narrow and closed environment of subway tunnels,lidar has significant advantages over traditional tunnel deformation detection methods.Due to the influence of the environment and the large number of equipment installed in the tunnel,the data will contain a lot of noise points,and the data of the complete inner wall of the tunnel cannot be obtained,thus affecting the application of the data in deformation detection and 3D modeling.This paper combined the least square method and Lagrange multiplier method to ellipse fit the tunnel point cloud model,and applied it to the circular tunnel point cloud filtering,and used the data of each region to detect the tunnel deformation.The accuracy evaluation and effectiveness analysis of the filtering results show that the algorithm has good filtering effect and stability for the tunnel point cloud data.