遥感影像库采样策略和空间配准快速算法研究  

Study on Sampling Schematism and Fast Algorithm of Spatial Registration for Remote Sensing Image Database

作　　者：郭晓英[1] 刘素红[1] 

机构地区：[1]北京师范大学地理学与遥感科学学院遥感与地理信息系统研究中心

出　　处：《计算机工程与应用》2005年第7期54-56,共3页Computer Engineering and Applications

基　　金：国家863高技术研究发展计划(编号:2002AA130010);国家重点基础研究发展规划项目(编号:G20000779)

摘　　要：建立面向应用的遥感影像库关键在于建立序列化数据的索引机制,空间上既要建立相应分辨率下的数据目标,同时又要反映不同分辨率之间数据的空间关系和组成结构。但是不同分辨率的遥感数据空间定位的策略不同,需要在空间采样和数据访问机制上进行关联。论文提出利用二进制剖分算法将不同空间分辨率的遥感影像数据进行多级自组织处理MLSO(Multiple-Layer Self-Organization),将中高低分辨率的数据根据相互关联的采样策略进行重采样处理。基于MLSO机制,论文开发了快速算法,在局域范围内进行相邻像元的搜索,快速实现不同分辨率影像数据序列化。经过实践证明,该方法快速、准确。Application-oriented remote sensing data need to be spatially serialized,that is,it must show spatial object in corresponding resolution as well as the spatial relationship and structure of various spatial objects in different resolutions.However,the spatial location strategies of remote sensing data in various resolutions are different and need to be associated with spatial sampling and data accessing.This story shows MLSO(Multiple-Layer self-Organization) of remote sensing data in different resolutions using binary section algorithm,re-sampling various resolution levels data according to their sampling strategies.Based on MLSO,this paper puts forward a quick algorithm,which can search neighbor pixels in local scope and serialize different resolutions image data rapidly.It is a quick and accurate algorithm proved by practical application.

关 键 词：遥感影像库 序列化 采样策略 快速算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]