基于图像内容层次表征的遥感图像分割方法  

Remotely sensed imagery segmentation based on the hierarchical representation of image content

作　　者：张学良[1] 肖鹏峰[1] 冯学智[1] 

机构地区：[1]南京大学地理信息科学系,南京210093

出　　处：《中国图象图形学报》2012年第1期142-149,共8页Journal of Image and Graphics

基　　金：国家高技术研究发展计划(863)基金项目(2008AA12Z106);国家自然科学基金项目(40801166);教育部高等学校博士学科点专项科研基金项目(200802841012)

摘　　要：提出基于图像内容层次表征的高分辨率遥感图像快速多精度分割方法。首先根据初始分割结果建立区域邻接图(RAG),并将其定义为马尔可夫随机场(MRF);然后引入光谱、形状和边缘等图像特征进行层次合并,通过记录层次合并过程获得图像内容的层次表征;最后根据层次表征中不同层级对象之间的关系快速生成任意不同精度的分割结果,以满足不同应用的需求。利用QuickBird卫星图像进行实验和评价的结果表明,本文方法具有较高的精度和效率。A segmentation method for high-spatial resolution remotely sensed images segmentation based on the hierarchical representation of the image content is proposed in this paper. The hierarchical representation is built through a hierarchical merging process based on an initial segmentation. First, we build the region adjacency graph （RAG） on which a Markov random field （MRF） is defined. Then the hierarchical merging process, where the merging criterion includes multi-spectral, edge and shape features, is applied and recorded on the RAG to obtain the hierarchical representation. It is quite efficient to produce a number of segmentation results with different precisions using the relationship of different objects in the hierarchical representation. Experimental results on QuickBird images show that the proposed method can greatly improve the segmentation efficiency and produce high quality results.

关 键 词：遥感图像分割 图像内容层次表征 层次合并 马尔可夫随机场 区域邻接图 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]