基于最小均方误差的圆弧分段曲线拟合方法  

METHOD OF CIRCULAR ARC FRAGMENTED CURVE-FITTING BASED ON LEAST MEAN-SQUARE ERROR

作　　者：涂嘉文[1] 徐守时[1] 谭勇[1] 

机构地区：[1]中国科学技术大学电子工程与信息科学系,安徽合肥230027

出　　处：《计算机应用》2001年第3期48-50,共3页journal of Computer Applications

摘　　要：提出了一种基于最小均方误差准则的圆弧分段曲线拟合方法。该方法采用自适应高斯滤波器对原始曲线进行平滑处理以消除噪声影响 ,并提出了一种适合于圆弧曲线拟合的分段算法。在该算法的基础上根据最小均方误差准则 ,以圆弧作为基元对曲线进行拟合。实验结果表明 ,该方法能够较好地消除噪声影响并保留曲线的局部特征。We present a method of circular arc fragmented curve-fitting based on least mean-square error rule in this paper. This method adopts self-adaptive gauss filter to process original curve smoothly to eliminate noise effects, and proposes a fragrmented algorithm that is adaptive to circular arc. Based on this algorithm, we use the circular arc as basic element to fit curve accroding to least mean-square error rule. The experiment results indicate that this method can eliminate noise effects as well as preserve the local characteristics of curve.

关 键 词：最小均方误差准则 高斯滤波器 圆弧分段曲线拟合 图形处理 计算机 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] O24[自动化与计算机技术—计算机科学与技术]