摘要
分水岭算法用于图像分割时通常由于噪声的存在和梯度的不规则造成过度分割。我们提出了一种新型的分水岭算法,结合了小波分析和数学形态学的一些概念,有效地减少了过度分割现象的产生。实验表明:本算法实现了预期的分割目标,取得了良好的效果。
Due to noise and irregularity of the gradient image, watershed algorithm used to segment images generally leads to over-segmentation, which is unacceptable. A novel watershed algorithm, which combines with some concepts of wavelet analysis and mathematic morphology, is proposed in this paper. The test shows that this method can solve the problem of over-segmentation and achieve the anticipative goal.
出处
《生物医学工程学杂志》
EI
CAS
CSCD
北大核心
2006年第5期991-994,共4页
Journal of Biomedical Engineering
关键词
图像分割
分水岭算法
过度分割
小波分析
数学形态学
Image segmentation Watershed algorithm Over-segmentation Wavelet analysis Mathematical morphology
