基于马尔科夫随机场的改变势函数的SAR图像分割  

An SAR Image Segmentation Based on Markov Random Field Variable Potential Function

作　　者：齐千慧 田益民[1] 陈红梅 韩利利 崔圆斌 杨益暄 QI Qian-hui;TIAN Yi-min;CHEN Hong-mei;HAN Li-li;CUI Yuan-bin;YANG Yi-xuan(Beijing Institute of Graphic Communication,Beijing 102600,China)

机构地区：[1]北京印刷学院,北京102600

出　　处：《电脑与信息技术》2021年第3期4-6,共3页Computer and Information Technology

基　　金：国家自然科学基金项目(项目编号:NSFC61178092);国家自然科学基金项目(项目编号:NSFC61378001);北京市教委科研计划资助项目(项目编号:KM202010015007)。

摘　　要：随着合成孔径雷达(Synthetic aperture radar)技术的发展,合成孔径雷达技术常应用于地表检测、灾害检测、环保检测工作等方面。SAR图像的分割处理也变得尤为重要。因为马尔科夫随机场(Markov Random Field)结合了邻域像素的空间相关性,所以基于MRF的图像分割法在SAR图像方面得到了广泛地应用。但由于该方法的势函数通常选用固定的经验值,导致了分割效果不佳。文章将邻域像素点的相关性引入到了势函数中,得到了一种基于马尔科夫随机场的改变势函数的SAR图像分割方法。With the development of SAR technology,SAR technology is often used in surface detection,disaster detection,environmental protection detection and so on.The segmentation and processing of SAR image becomes particularly important.Because Markov random field combines the spatial correlation of neighborhood pixels,MRF based image segmentation method has been widely used in SAR images.However,because the potential function of this method usually USES fixed empirical values,the segmentation effect is not good.In this paper,the correlation of neighborhood pixels is introduced into the potential function,and a SAR image segmentation method based on Markov random field variable potential function is obtained.

关 键 词：数字图像处理 马尔科夫随机场 图像分割 SAR图像 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]