径向基网络的非线性插值与3D点云曲面重构技术  

Nonlinear Interpolation of Radial Basis Networks and 3D Point Cloud Surface Reconstruction Technique

作　　者：陈文兴 闫丽萍 崔英 李苗 CHEN Wen-xing;YAN Li-ping;CUI Ying;LI Miao(School of Mathematics and Statistics,Ningxia University,Yinchuan 750021,China;School of Mathematics and Statistics,Xidian University,Xi'an 710126,China)

机构地区：[1]宁夏大学数学统计学院,宁夏银川750021 [2]西安电子科技大学数学统计学院,陕西西安710126

出　　处：《数学的实践与认识》2019年第1期282-287,共6页Mathematics in Practice and Theory

基　　金：宁夏大学研究生创新项目(GIP2018069)

摘　　要：研究了径向基网络插值算法与3D曲面重构方法,分别从研究价值、插值理论、仿真等等方面做了详细分析,研究结果表明RBF-Network在逼近一维复杂的非线性函数时具有收敛速度快、精度高、泛化能力更强等优点.但在3D重构方面,基于RBF的单位分解法重构效果更好.This paper studies the radial basis network interpolation algorithm and 3 D surface reconstruction method. We analyzed the research value, interpolation theory, simulati-oion and so on. The research results show that RBF-Network has the advantages of fast convergence, high precision and generalization ability when approaching one dim ensional complex nonlinear functions. However, in terms of 3 D reconstruction, the partition of unity method based on RBF is more effective.

关 键 词：非线性插值 泛化误差 Delauney三角 点云插值 3D曲面重构 

分 类 号：O241.3[理学—计算数学]