基于图像配准的复杂背景下自动目标识别算法  

An algorithm based on image registration for automatic target recognition in complex environment

作　　者：王明静[1] 宋玲珍[2] 

机构地区：[1]中国空空导弹研究院,河南洛阳471009 [2]洛阳师范学院,河南洛阳471022

出　　处：《中国体视学与图像分析》2009年第3期266-270,共5页Chinese Journal of Stereology and Image Analysis

摘　　要：提出了一种基于图像配准的自动目标识别算法,图像配准算法采用基于归一化互信息相似性判据,并采用模糊自适应粒子群优化算法作为搜索策略。在图像精确配准的基础上,通过图像间的相互转换,间接实现了目标的准确识别。仿真试验结果表明,该方法可以实现复杂背景下目标的准确识别。An automatic target recognition algorithm based on image registration is presented.The image registration uses fuzzy adaptive particle swam optimization（PSO）algorithm as the search strategy,and the normalized mutual Information is used as the similarity criteria.Based on the accurate registered result,target recognition is correctly achieved indirectly through the transforming between the images.Experimental result s show that the targets in complex environment can be recognized correctly.

关 键 词：自动目标识别 图像配准 归一化互信息 模糊自适应PSO 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]