摘要
提出了一种将基于群体增量学习(population-based incremental learning,PBIL)算法用于遥感影像匹配的方法,给出了详细的理论和实验分析,引入信息熵作为PBIL算法迭代终止的条件之一,取得了较好的实验结果。实验表明,基于该算法的影像匹配运算速度比较快,且收敛过程比较稳定。
A PBIL algorithm is presented to match remote sensing images. It uses the strategies of genetic operation and competitive learning, modifies the learning probabilities according to competitive learning, and then supervises the offspring generation. The detailed academic and experimental analysis are introduced, and the information entropy as one of the iterative terminated conditions is put forward. Experimental results show that this method is effective and fast, and the convergence procedure is stable.
出处
《武汉大学学报(信息科学版)》
EI
CSCD
北大核心
2008年第2期140-143,共4页
Geomatics and Information Science of Wuhan University
基金
国家十一五预研资助项目
关键词
PBIL算法
影像匹配
信息熵
PBIL algorithm
image matching
information entropy