基于主动轮廓模型的SAR图像分割方法综述  

Overview of SAR image segmentation based on active contour model

作　　者：涂松[1] 李禹[1] 粟毅[1] 

机构地区：[1]国防科学技术大学电子科学与工程学院,湖南长沙410073

出　　处：《系统工程与电子技术》2015年第8期1754-1766,共13页Systems Engineering and Electronics

基　　金：国家自然科学基金(61171135)资助课题

摘　　要：针对合成孔径雷达(synthetic aperture radar,SAR)图像分割这一研究热点,综合论述了基于主动轮廓模型(active contour model,ACM)的SAR图像分割方法。首先,介绍了经典的ACM及其数学原理,并通过理论和实验分析了这些模型应用于SAR图像分割时存在的问题;然后,对目前基于ACM的SAR图像分割方法进行了系统的梳理和分类讨论;最后,对基于ACM的SAR图像分割方法作了总结,并对将来的研究方向进行了展望。Recently, a great number of interests have been attracted to image segmentation methods based on the active contour model （ACM）. The classical ACM and synthetic aperture radar （SAR） image segmenta- tion methods based on the ACM are reviewed. Firstly, classical ACMs and mathematical principles of them are introduced, and the defects are analyzed when they are applied to SAR image segmentation. Secondly, classified discussion and analysis of SAR image segmentation methods based on the ACM are given. Finally, several fu- ture research directions for SAR image segmentations based on the ACM are discussed.

关 键 词：合成孔径雷达图像分割 主动轮廓模型 统计建模 凸优化 相似性度量 

分 类 号：TN957.52[电子电信—信号与信息处理]