摘要
本文的主要工作有以下三个方面:首先通过多方面分析,指出文献中Hough变换的模板匹配式定义是不完善的;其次我们发现Hough变换具有一种固有的概率特征,它和待处理的图像无关,但对Hough变换的性能有根本性的影响,为了描述这种概率特征,本文提出了曲线(曲面)密度函数的概念;最后基于上述两方面的分析,本文给出了Hough变换的一种新定义.
This paper mainly contributes the following 3-fold. Firstly, it is shownthat the template matching-like definitions of the present Hough transform are inadequate. Secondly, an inherent probabilistic aspect of the Hough transform is clarified, and a new term, namely curve (surface) density function, is introduced tocope with the undesirable effects coming from such a probabilistic aspect. Finally,a new definition of the Hough transform is proposed, which includes mainly theused parameterization and associated curve (surface) density function.
出处
《计算机学报》
EI
CSCD
北大核心
1997年第8期744-752,共9页
Chinese Journal of Computers
基金
国家自然科学基金
国家863高科技基金
关键词
HOUGH变换
参数化
模板匹配
图像处理
Hough transform, parameterization, curve (surface) density function.