针对观测向量和系数矩阵均含有误差以及点云数据存在异常点的问题,该文提出一种稳健加权总体最小二乘法。该方法在加权总体最小二乘的基础上,通过设置一定的准则,剔除点云数据中存在的异常点,以获取更为精确的平面拟合参数解。仿真模拟算例和实际点云数据实验结果表明,该方法与传统的方法相比,能够消除异常点带来的影响,获得更精确的参数解,平面拟合精度更高。
In this paper, a method of robust weighted total least squares was proposed. Based on weighted total least squares and certain criteria, the outliers in point clouds were detected and removed to obtain more accurate parameter solution of plane fitting. The experiments result of plane fitting with simulated point clouds and real point clouds showed that comparing with traditional methods, the proposed method could eliminate the effects caused by outliers, obtain more accurate parameter solution and better accuracy of plane fitting.