基于三维激光点云的地铁隧道特征提取方法研究

Research on the Method of Feature Extraction in Subway TunnelsBased on Three-dimensional Laser Point Clouds

针对地铁隧道的结构化特点,为实现轨道及月台平面的自动化巡查,本文提出一种基于几何约束的轨道及平面点云特征快速提取方法。对于隧道激光点云,通过PCA估计的方法确定其点云主轴方向并将其矫正至水平;在此基础上利用直通滤波提取隧道横截面点云,通过RANSAC拟合构建横截面数学模型;然后计算截面上各点到拟合圆心的距离,依据点心距的变化提取候选轨道点与平面点;根据两组截面的特征点优化求解直线与平面方程,计算点线、点面距离,完成直线、平面的特征提取。研究结果表明,轨道检测精度较好,检测准确率高。

In view of the structural characteristics of subway tunnels,in order to realize automated inspection of tracks and platform planes,this paper proposes a rapid extraction method of track and plane point cloud features based on geometric constraints.First,for the tunnel laser point cloud,the direction of the main axis of the point cloud is determined through PCA estimation and corrected to the level;on this basis,straight-through filtering is used to extract the tunnel cross-section point cloud,and the cross-section mathemat-ical model is constructed through RANSAC fitting;secondly,it calculates the distance from each plane point on the section to the cen-ter of the fitting circle,and extracts candidate orbit points and plane points based on the change in point-center distance;finally,it optimizes and solves the straight line and plane equations based on the characteristic points of the two groups of sections,calculates the point-line and point-surface distances to complete the straight line and planar feature extraction.The research results indicate that the orbit detection accuracy is good and the detection accuracy is high.