边缘检测中小波选取问题研究  

Study of wavelet selection for edge detection based on wavelet’s self-characteristics

作　　者：王慧燕[1] 程翼宇[2] 

机构地区：[1]浙江工商大学计算机与信息工程学院,浙江杭州310018 [2]浙江大学药物信息学研究所,浙江杭州310013

出　　处：《电路与系统学报》2006年第3期48-53,共6页Journal of Circuits and Systems

基　　金：国家中医药管理局重点项目(2000-J-Z-03);浙江省科技计划项目(011101066)

摘　　要：小波的选取对边缘检测结果有直接的影响。本文分析和研究了小波的对称性、收敛性、平滑性及小波消失矩的内在联系及这些特性对边缘检测的影响,提出了边缘检测中小波选取原则。实验选取了三种具有不同特性的典型小波对选取原则进行了验证,实验结果证明了本文提出的小波选取原则的正确性。In the methodology of edge detection based on wavelet transform, the wavelet function is always selected by experiments, which is blind and time-consuming. In this paper, we studies the wavelet＇s characteristics including symmetry, convergence, regularity and wavelet cancellations and their influences on edge detection and proposes wavelet selection principles in edge detection. Three typical wavelets with different characteristics are tested to validate the selection principles The experiment results demonstrate that the proposed principles are accurate and can be used to direct the selection of wavelet in edge detection.

关 键 词：小波选取 边缘检测 收敛性 对称性 平滑性 消失矩 

分 类 号：O231.2[理学—运筹学与控制论]