基于目标识别的红外与微光图像融合方法  

Fusion of infrared and LLL images based on dynamic target recognition

作　　者：张俊举[1] 崔东旭[1,2] 贺峻峰[2] 聂伟乐[2] 韩亦勇[1] 袁轶慧[1] 常本康[1] 

机构地区：[1]南京理工大学电子工程与光电技术学院,江苏南京210094 [2]西安应用光学研究所,陕西西安710065

出　　处：《应用光学》2011年第6期1134-1138,共5页Journal of Applied Optics

摘　　要：为了在融合图像中突出运动目标,提出了一种基于动态目标检测和识别的图像融合算法。先对红外图像序列中的运动目标进行检测和提取,同时对红外和微光图像进行融合,最后将提取到的红外目标与融合图像进行二次融合。试验结果表明,该算法获得的融合图像不仅具有普通融合算法信息丰富的特点,还具有鲜明的红外目标指示特性。To highlight the moving targets in fusion images,an imaging fusion algorithm based on target recognition is presented.Firstly,moving targets are detected from dynamic infrared images.At the same time,infrared and low-light-level（LLL） images are fused with the contrast pyramid method.Finally,the profile images of moving targets extracted from infrared image sequences are fused into the fusion image sequences.The experimental results show that the fused images obtained by the algorithm provides rich information similar to conventional fusion methods and enhances target signatures.

关 键 词：夜视技术 多光谱成像 图像融合 动态目标识别 

分 类 号：TN223[电子电信—物理电子学]