曲线曲面小波分解偏差估计  

Deviation Estimating of Curves and Surfaces in Wavelet Transform

作　　者：俞佳庆[1,2] 赵罡[1,2] 

机构地区：[1]北京航空航天大学机械工程及自动化学院,北京100191 [2]北京航空航天大学虚拟现实技术与系统国家重点实验室,北京100191

出　　处：《计算机辅助设计与图形学学报》2012年第8期984-988,996,共6页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家自然科学基金(61170198)

摘　　要：针对曲线曲面小波变换后低分辨模型和原始高分辨模型之间的偏差问题,提出离散偏差模型方法来估计偏差.通过定义偏差曲线,并在此基础上提出离散偏差曲线法来估计高分辨曲线和低分辨曲线之间的偏差:给定弓高误差限,用等误差方法离散偏差曲线,利用离散点估计偏差.同时,将该思想推广到曲面小波分解偏差估计上,定义了偏差曲面,并提出离散偏差曲面法来估计高分辨曲面和低分辨曲面之间的偏差.实验结果表明,文中算法既可以实现分解n次之后的低分辨模型的偏差估计,又可以通过减小离散误差提高估计精度.In this paper, deviation model discrete methods are presented to estimate the deviation between the low and high resolution curves and surfaces in wavelet transform. First, deviation curve is defined, and the deviation curve discrete method is presented： discrete the curve by the arc error bound, and then using the discrete points to estimate the deviation. Meanwhile, the curve method is also extended to surface deviation estimating. Deviation surface is defined, and the deviation surface discrete method is presented to estimate the deviation of low resolution surfaces. The experimental results demonstrate that the methods in this paper not only can estimate the deviation between the high resolution model and the low resolution model which are after n times wavelet transform, but also can improve the accuracy by reducing the error bound during model discrete.

关 键 词：小波变换 偏差估计 曲线曲面离散 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]