摘要
在用图像增强、细化技术对红外图像进行预处理基础上,提出将边缘轮廓图像的特征分为交叉点、分支点和一般点,并对这三类点集赋予不同权值的计算方法及使用蚁群智能寻优Hausdorff距离,提高了红外图像匹配算法的鲁棒性及计算性能。仿真验证了本文算法的有效性。
Based on the strengthening of images and thinning technology to pre-process the infrared image, The characteristics of verge outline images can be divided into cross point, bifurcated point and common point .The calculating method of giving three kinds endows different fight and using ant colony brainpower seeks optimal path and improves the matching robust of infrared image and the performance of calculating. Simulating proves the validity of this algofithm.
出处
《红外技术》
CSCD
北大核心
2007年第12期708-711,共4页
Infrared Technology
基金
航空科学基金(04I53067)
航空支撑科技基金(05C53005)
关键词
红外图像
蚁群算法
图像匹配
infrared image
ant colony algorithm
matching of image
