基于激光点云的隧道断面自适应提取方法

Self-adaptive Profile Extraction of Tunnel Based on Laser Point Cloud

隧道作为狭长型密闭工程建筑,在其全生命周期都需要进行安全检测。采用三维激光扫描技术获取的海量点云数据可更好地反映隧道的整体信息。但因扫描设站距离不同以及遮挡等情况,点云密度不均匀,不利于后续的点云特征提取及形变分析。针对这个问题,该文提出了一种能够自适应地调整断面厚度从而提取不同里程处断面点云的方法,然后对提取断面点云进行降噪以及三次B样条曲线拟合,并将拟合结果分别与贝赛尔曲线、拉格朗日曲线进行了对比。实验表明,该方法提取断面精度可达毫米级,能够较好地反映隧道真实变形情况。

As a long and narrow closed engineering construction,the tunnel needs be detected in its whole life cycle.The massive point cloud data obtained by 3D laser scanning technology can better reflect the total deformation information of tunnel.But the point cloud density is uneven due to the different distance of stations and occlusion,which is difficult to extract the subsequent point cloud characteristic and deformation analysis.Given to the problem,this paper proposes a method that can adaptively adjusts the section thickness to extract the point cloud of different mileage sections.And then the extracted section point cloud is denoised and fit the B-spline curve three times.The fitting results are compared respectively with Bezier curve,Lagrange curves.The experimental results show that the accuracy by this method can reach to scale of milli-meter,which can reflect the real deformation of tunnel.