基于特征的多层次红外与可见光图像融合方法  

Infrared and visible image fusion based on multilayer feature extraction

作　　者：张宇[1] 付冬梅[1] 李晓刚[1] 孔维功[1] 

机构地区：[1]北京科技大学信息工程学院,北京100083

出　　处：《计算机应用研究》2009年第4期1549-1551,共3页Application Research of Computers

基　　金：国家自然科学基金资助项目(60573016);北京市教委重点学科共建项目(XK100080431)

摘　　要：目前基于融合的方法能够改善红外图像的视觉效果,但局限于简单的直接融合,忽略了背景等因素所含噪声的影响及各部分细节信息。对此做了进一步研究工作,改进了现有方法的融合规则,提出先将目标从背景中提取出来再以温度阈值为依据分层融合,并加入实时温度信息,可随时得到融合图像的温度数据,从而在细节上极大地改善了目标的视觉效果及信息含量,提高了效率;最后对融合效果进行了定量评价和比较。实验结果证明处理后的图像能够比原图像获得更丰富的视觉信息。Though the presented method of fusion can improve the visual effect of infrared image, it is still confined to single and direct fusion which neglect the influence of background＇ s yawp factor. The paper made farther research on the respect, im- proved the existence fusion＇ s regulation through proposed a method that firstly extracted the objective, then fused the images layered which based on temperature threshold, and joined temperature of the infrared image in the fusion result real-time. Thereby improved the specific of visual and effect largely. Finally, evaluated the effect of the result through objective method. The result shows that the method can obtain more information than the original image.

关 键 词：可见光图像 红外图像 目标提取 温度阈值 分层融合 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]