摘要
传统的遥感图像过渡区提取与分割算法基于梯度运算。指出了传统算法存在的不足,对遥感图像过渡区像素属性进行分析,使用局部复杂度参数用于遥感图像的过渡区提取与分割。实验研究表明,算法优于传统的基于梯度算子的过渡区提取与分割方法,算法抗噪性好,鲁棒性强。
Traditional remote image transition region extraction and segmentation methods are on the basis of gradient operator. Their drawbacks are discussed at first. By analyzing properties of transition regions, a novel local complexity based transition region extraction method (C-TREM) is proposed for remote image segmentation. Experimental results demonstrate that C-TREM significantly outperforms the conventional gradient-based transition region extraction methods (G-TREM). C-TREM has good ability to deal with noises. It is a robust method for remote image transition region extraction and segmentation.
出处
《武汉理工大学学报》
CAS
CSCD
北大核心
2009年第18期160-163,共4页
Journal of Wuhan University of Technology
关键词
图像分割
局部复杂度
过渡区
遥感图像
image segmentation
local complexity
transition region
remote image