景象匹配算法研究进展与展望  

Status and Prospect of Algorithm for Scene Matching Systems

作　　者：谭志园[1] 孙继银[2] 王忠 张财兴[1] 

机构地区：[1]中国人民解放军96401部队 [2]第二炮兵工程大学

出　　处：《计算机技术与发展》2012年第9期66-70,共5页Computer Technology and Development

基　　金：总装武器装备预研基金(513220208)

摘　　要：景象匹配是精确制导武器精确定位目标的重要方法,景象匹配算法决定了精确制导武器系统的作战性能。论文介绍了景象匹配算法实现的关键要素,详细综述了具有代表性的基于区域、基于特征、基于变换域以及基于模型的四类匹配算法的不同原理、适用性及其改进算法,探讨了景象匹配算法中有待进一步研究的问题和未来发展的方向,为未来的景象匹配制导研究工作提供技术参考和依据。景象匹配制导技术正朝着更稳定可靠、更智能、更快捷、更精确的方向发展,必将推动精确制导武器装备实现现代化和自动化。Scene matching is the important method to pinpoint targets for precision guide weapon. Scene matching algorithm determines the operational performance of the precision weapons system. In this paper,the key elements of scene matching algorithm are introduced, describe the different principle and applicability of some representative algorithms which include the classical algorithms based on region, feature, transformation domain and model as well as its improved method in detail, the issue needed to be investigated farther and future direction is discussed. Provide the technical reference and basis for scene matching systems in the future. The development of scene matc- hing guidance technology which will push precision weapons equipment minute and automatic will be more intelligent,faster and more accurate.

关 键 词：景象匹配 相似性度量 矩不变量 角点 SIFT 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]