摘要
为了解决图像匹配过程中计算速度慢和匹配精度不高的缺陷,提出了一种基于群体增量学习(Population-based Increased Learning,简称PBIL)算法的匹配方法。PBIL算法是一种基于概率分析的进化算法。它集成了基于函数优化的遗传搜索和竞争学习两种策略,将进化过程视为学习过程,通过竞争学习所获得知识来修正生成概率,进而指导后代的生成。给出了理论分析和实验分析。在实验中,分析了不同终止条件下的算法性能,并将其与传统序贯相似性检测算法(SSDA)和遗传算法进行了比较。实验结果表明基于该算法的图像匹配具有运算速度快、匹配精确等优点,且收敛过程非常稳定。
To solve the problem of slow computation speed and low accuracy in image matching,a new approach to image matching using Population-based Increased Learning algorithm（PBIL） has been proposed.PBIL algorithm is a probability learning based Evolutionary Algorithm.It integrates genetic search strategy based on function optimization with competitive learning strategy,regards evolution as a learning process,revises the generation probability according to knowledge come from competitive learning,then produces the offspring according to the probability.Theoretic analysis and experimental analysis are presented.In experiment,have analyzed the algorithmic performance of different finish conditions,and then compared it with the conventional sequential similarity detection algorithm and genetic algorithm. Results of experiment show this approach is fast in operation,and has high accuracy in matching,and the convergence is very stable.
出处
《计算机工程与应用》
CSCD
北大核心
2005年第25期43-45,87,共4页
Computer Engineering and Applications
基金
国家科技成果重点推广项目计划(编号:2004EC000096)资助
关键词
PBIL算法
图像匹配
相关匹配
遗传算法
PBIL algorithm,image matching,correlation matching, Genetic Algorithm
