摘要
本文基于数学形态学理论提出了描述物体形状特征的形态谱分量函数并讨论了其性质。该函数能够给出难以用经典的形态谱描述的形状方向信息，并能有效地应用于形状分析。给出了实际图象提取形态谱分量函数的方法，提出了有效的距离函数并用于形状分析。
In this paper a pattern spectrum component function is introduced and tis properties are discussed It can represent directional information of shape rnore precisely than the oriented pattern spectrum proposed by P.Maragos, and can be properly used in shape analysis. An appropriate distance function between two pattern spectrum functions is presented to recognize different shapes in noise. As a practical application, the appropriate distance function is applied to warning sign recognition and then favorable results are obtained.
出处
《信号处理》
CSCD
1998年第4期291-295,共5页
Journal of Signal Processing
基金
国家自然科学基金
关键词
形状分析
数学形态学
形态谱分量函数
图象识别
Shape analysis, Mathematical morphology, Oriented pattern spectrum,Pattern spectrum component function, Distance function
