A New Visible and Infrared Image Fusion Algorithm Based on NSCT

Abstract

This paper propose an image fusion algorithm for infrared (IR) and visual (VIS) image. The proposed method firstly decompose input images into low and high frequency sub-bands using Non-Subsampled Contourlet Transform (NSCT). Secondly, an adaptive weighted fusion rule is developed to fuse the low frequency sub-band of input image. The proposed fusion rule generate weight map using Gaussian function in low frequency sub-band. To highlight the high frequency feature, the high frequency sub-bands is fused by the max absolute rule. Finally, the fused image is generated using the fused sub-band images by the inverse NSCT. The experiments show that the proposed method provides improved subjective visual and objective evaluation results compared with GFL and average fusion rules.