摘要
本文基于图象的分层表示,提出了一种边缘检测的分层松弛方法。实验结果表明,本文方法计算简单,运算速度快,而且不受噪声影响,能有效地应用于细胞图象的边缘检测。
A new hierarchical relaxation method for edge detection is proposed. It is shown that using multiresolution information of the image hierarchical representation, the weighted coefficient in the iteration process can be selected by taking into account the values at the high level of the pyramids as a measure of the local edge features in the images. Satisfactory detection can then be obtained at all points of the images, and the eonvergence rate of the relaxation is greatly improved. This method is also useful for obtaining the complete border of the objects in the images.
出处
《华中理工大学学报》
CSCD
北大核心
1990年第3期139-144,共6页
Journal of Huazhong University of Science and Technology
关键词
图象处理
边缘检测
特征提取
Image processing
Edge detection
Feature extraction
Relaxation method
Hierarchical representation
Pyramids
