一种改进的自适应遗传算法在影像匹配中的应用  

Application of an improved adaptive genetic algorithm in image matching

作　　者：李国[1,2] 范大昭[1] 郭海涛[1] 

机构地区：[1]信息工程大学测绘学院 [2]65015部队

出　　处：《测绘科学》2009年第5期188-189,80,共3页Science of Surveying and Mapping

基　　金：信息工程大学测绘学院第二届学员科技创新基金项目(X0608)

摘　　要：影像匹配是计算机视觉及数字摄影测量的核心问题。基于遗传算法的全局寻优特点及影像匹配的具体特点,本文将遗传算法引入到遥感影像的匹配定位中。针对遗传算法中变异概率、交叉概率在实验中始终保持不变的缺点,本文利用Fibonacci数列对基于遗传算法的影像匹配进行了自适应改进。实验结果表明这种方法可以产生一个快速而且可靠的匹配结果。Image matching is a kernel problem in computer vision and digital photogrammetry. Based on the excellent characteristic of seeking for global optimization and specific characteristic of image matching, genetic algorithm is introduced in the matching and location of sensor images. The probabilities of crossover and mutation are invariability, so an adaptive genetic algorithm with Fibonaeei sequence was realized in image matching. The results show that the method discussed can produce a reliable matching result in high speed.

关 键 词：影像匹配 遗传算法 FIBONACCI数列 GRAY码 

分 类 号：P23[天文地球—摄影测量与遥感]