摘要
异类图(不同传感器所成的图像)的匹配具有重要性和挑战性,与传统的匹配相比有很大不同。从可见光、红外以及雷达的成像机理出发,对异类图匹配的方法进行了分析和总结,对其适应性和局限性进行了归纳,提出了异类图匹配难点和研究方向,对异类图的匹配研究具有一定的指导和借鉴意义。
Difference between the conventional image registration and multisensor image registration (images acquired by different sensors) is an important and difficult problem. Originated from the basic principles of different images such as optical image, infrared image and radar image, several algorithms of multisensor image registration are analyzed and summarized. The adaptability and limitation of these algorithms are also presented. Some important problems and difficulties are pointed out, which may be valuabl...
出处
《红外与激光工程》
EI
CSCD
北大核心
2008年第S2期646-650,共5页
Infrared and Laser Engineering
基金
国家自然科学基金(60574084)
国家973计划(2002CB312200)
关键词
异类图
景象匹配
直方图规定化
边缘检测
互信息
Multisensor images
Image registration
Histogram specification
Edge extraction
Mutual information
