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

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.