摘要
Otsu分割算法是阈值分割技术中广泛使用的一类方法,但它不能准确分割对比度较低或变化较大的图像。借助于数学形态学中高帽变换的概念,提出了一种改进的Otsu算法:先用高帽变换对原始图像处理,在处理后的图像上再进行Otsu分割。仿真实验表明,该方法具有较低的算法复杂度,容易实现且能成功分割出Otsu法不能分割的对象,比Otsu法具有更好的鲁棒性。
Otsu segmentation algorithm is used for image threshold segmentation widely, but it can＇t segment images with low or variable contrast exactly. By the help of concept of Top-hat transformation in mathematical morphology, a modified Otsu method is proposed, by which the original images are processed with Top-hat transformation firstly and then segmented by Otsu method. The results of simulation experiments show that the method has advantages of low algorithmic complexity , easy realization, and can segment the special images which Otsu can＇s deal with effectively. It has more perfect robustness than Otsu
出处
《传感器世界》
2011年第12期9-11,共3页
Sensor World
基金
山西省青年科技研究基金资助项目(项目编号:2009021019-2)
关键词
高帽变换
OTSU法
阈值分割
Top-hat transformation
Otsu method
threshold segmentation
