面向遥感影像感兴趣目标的剖分表达  

Remote Sensing Images Objects of Interests Subdivision Expression

作　　者：宋树华[1,2] 程承旗[1] 濮国梁[1] 郭仕德[1] 官小波 

机构地区：[1]北京大学遥感与地理信息系统研究所,北京100871 [2]中国资源卫星应用中心,北京100094 [3]中国农业银行资产负债管理部,北京100005

出　　处：《北京工业大学学报》2013年第11期1652-1659,共8页Journal of Beijing University of Technology

基　　金：国家"973"计划资助项目(61399)

摘　　要：为了高效地表达遥感影像中感兴趣目标,提出了基于地图分幅拓展EMD剖分面片聚合的遥感影像感兴趣目标剖分表达模型以及相应的剖分表达原则和剖分表达流程,设计了基于感兴趣目标剖分表达模型的线状和面状目标边界表达算法、面状目标的行扫描算法和剖分面片聚合算法等.模拟验证结果表明:基于EMD遥感影像感兴趣目标剖分表达模型可有效地对影像中的目标进行表达,且表达感兴趣目标的剖分面片个数越多,剖分面片聚合层级越大,目标剖分面片的压缩效果越好.To display objects of interest in remote sensing images （RSImg-OOIs） efficiently, RSImg- OOIs subdivision expression model based on Extended Mapping-Division （EMD） and subdivision expression principles and processes were proposed. It also illustrated algorithms, including linear object and planar object＇s boundary subdivision and expression algorithm, planar object line-scanning filling algorithm and subdivision cells aggregation algorithm. At last, the model and algorithms were tested by a simulation system. Results show that the subdivision expression model of RSImg-OOIs based on EMD is feasible to display RSImg-OOIs and the count of RSImg-OOIs subdivision cells can be greatly compressed by using the algorithm of subdivision cells aggregation with increasing the number of subdivision cells and the levels between aggregation cells.

关 键 词：遥感影像感兴趣目标 全球剖分模型 地图分幅拓展 剖分面片 

分 类 号：P208[天文地球—地图制图学与地理信息工程]