LWIR and MWIR Images Fusion Method Based on the TV Model and Saliency Analysis

Medium and Long wave infrared images are widely used in maritime target detection. However, the target detection effect is seriously affected by noise due to the incompleteness and blurry edges of the target in the image. Combined with the total variation model and significance analysis, we proposed a NSCT-based LWIR and MWIR image fusion method, to improve the salience of the target and detail information in the fusion image. The TV model maintains the same LWIR grayscale and preserve the coarse texture of MWIR in fusion images. Saliency analysis construct the weight map which retains the significant fine details of LWIR and MWIR images in the final result. Experimental results indicate that our method outperform other three commonly used method both in visual and evaluation metric aspects.