摘要
传统Otsu阈值法对图像分割缺乏一定鲁棒性的不足,将Otsu阈值法的平方欧氏距离用绝对值距离代替得到一种新的图像阈值化分割准则,提出了基于绝对值距离的图像阈值分割新算法,最后给出其参数估计方法。实验结果表明,分割新算方法是有效的,能够获得比传统Otsu方法更好的分割结果。
Considering that the classical Otsu＇ s thresholding method lacks robustness in image segmentation,in this paper we replace the squared Euclidean distance in Otsu＇ s thresholding segmentation criterion with the absolute distance to obtain a new image thresholding seg- mentation criterion, thus a new image thresholding segmentation algorithm based on absolute distance is presented. In the end of the paper, the parameter estimation method of this algorithm is introduced. Experimental results show that the segmentation method presented in the paper is effective and performs better in image segmentation than the classical Otsu＇ s segmentation method.
出处
《计算机应用与软件》
CSCD
2010年第8期259-262,共4页
Computer Applications and Software
关键词
图像分割
阈值化方法
OTSU法
平方欧氏距离
绝对值距离
Image segmentation Thresholding method Otsu＇s method Squared Euclidean distance Absolute distance
