论文部分内容阅读
随着三维激光技术的广泛应用,点云数据后处理技术直接影响到后期建模质量。其中在使用曲面拟合进行三维激光扫描点云数据滤波时,通常面临阈值选择或多次迭代的问题。为此,本文以Gauss-Seidel迭代为基本思想,移动最小二乘曲面拟合与Lagrange算子相结合的插值方法进行滤波,针对该算法处理过程中遇到方程组解无实根的情况提出解决方案,并对处理后的数据进行对比以验证算法有效性。结果表明,方差为0.5的随机噪声滤波效率达到80%,方差为1的随机噪声滤波效率达到90%。
With the wide application of 3D laser technology, point cloud data post-processing technology has a direct impact on the quality of post-modeling. Among them, the use of surface fitting for three-dimensional laser scanning point cloud data filtering, usually faced with the problem of threshold selection or multiple iterations. Therefore, in this paper, the Gauss-Seidel iteration as the basic idea, the moving least square curve fitting and Lagrange interpolation algorithm is used to filter, and the solution to the problem that the solution of the system of equations is unreal root Program, and compare the processed data to verify the validity of the algorithm. The results show that the random noise with the variance of 0.5 is 80% and the random noise with the variance of 1 is 90%.