论文部分内容阅读
图像分割是SAR图像分析的关键步骤,高质量的分割结果是发挥SAR图像应用潜力的保证.近年来,在SAR传感器蓬勃发展的背景下,SAR图像应用受到更为广泛的重视.然而相比光学图像,SAR自身的成像特点使其在分割时更难得到理想的结果.而随着模式识别、机器学习、遥感技术等相关学科领域的发展,SAR图像分割研究取得了快速进展.系统地总结了SAR图像分割的相关研究进展,在归纳分类的基础上,重点对基于模糊C均值、马尔科夫随机场、区域信息、统计分布、水平集、多层次特征及深度学习等热点方法的发展与最新研究进行了综述.最后针对SAR图像分割技术进行了展望.“,”Image segmentation is an indispensable part for synthetic aperture radar (SAR)image automatic interpretation.Segmentation results with high-quality are the guarantee of SAR image applications.In addi-tion,owing to the development of SAR sensors,the segmentation task based on SAR image has been wide-ly concerned in recent years.However,compared with the optical images,the unique properties of SAR ima-ges lead to great challenge in SAR image segmentation.With the development of pattern recognition,ma-chine learning,remote sensing technology and other related techniques,SAR image segmentation has made great progress.This paper reviews the progress of SAR image segmentation,and then puts its emphasis on the summary of the widely used algorithms:FCM,MRF,statistical model,region information,level set, multi-scale and deep learning,etc.Finally,several viewpoints for the future research of SAR image segmen-tation are proposed.