An Algorithm of Infrared and Visible Light Images Fusion Based on Infrared Object Extraction

Abstract

The traditional image fusion methods of Infrared and visible light images neglect differences between the background and the targets, resulting in the poor clarity or week identifiability of the fused image. According to the characters of infrared and visible images, an algorithm based on object extraction is proposed. This algorithm infrared image object is extracted by maximum between-cluster variance, the acquired object information and the background and detail information of visible light image are fusion by NSCT. Thereby improved the specific of visual and effect largely. Finally, the effect of the results is evaluated through objective method. Experimental results indicate that the image fusion obtained by this algorithm possessed the same object as the one in the infrared image, and keeps the same details information as the one in the visible image.