Fast algorithm of the robust Gaussian regression filter for areal surface analysis
Fast algorithm of the robust Gaussian regression filter for areal surface analysis
复制标题
DOI：
10.1088/0957-0233/21/5/055108
复制
发表时间：
2010-05
影响因子：
2.4
通讯作者：
W. Zeng;Xiangqian Jiang;Paul J. Scott;Paul J. Scott
中科院分区：
文献类型：
--
作者：
W. Zeng;Xiangqian Jiang;Paul J. Scott;Paul J. Scott
本文探讨了用于面积表面分析的高斯回归滤波器的通用模型。解决了线性高斯滤波器和鲁棒滤波器之间的内在关系。提出了该模型的一般数学解决方案。基于该技术，创建了一种快速算法。模拟和实际工程数据（随机和结构化）都已用于快速算法的测试。结果表明，在相同精度下，二阶非线性回归滤波器对1024*1024点数据集的处理时间由传统算法的数小时缩短至数秒。
In this paper, the general model of the Gaussian regression filter for areal surface analysis is explored. The intrinsic relationships between the linear Gaussian filter and the robust filter are addressed. A general mathematical solution for this model is presented. Based on this technique, a fast algorithm is created. Both simulated and practical engineering data (stochastic and structured) have been used in the testing of the fast algorithm. Results show that with the same accuracy, the processing time of the second-order nonlinear regression filters for a dataset of 1024*1024 points has been reduced to several seconds from the several hours of traditional algorithms.