Rolling Guide Filtering and Non-subsampled Contourlet Transform for Fusion of Visible and Infrared Images

Abstract

ABSTRACT Image fusion is essentially an image enhancement technology, which aims to generate fusion images with richer information and more features by extracting complementary information from images collected by different sensors (such as infrared and visible light) or the same sensor (such as multi-focus image). For the fusion of infrared and visible images, it is easy to produce problems such as missing detail information and suppressing less noise. In this paper, we propose a Visible and Infrared Images fusion method by combining Non-subsampled Contourlet Transform (NSCT) and rolling guide filtering. First, fuzzy logic algorithm is used to enhance the contrast of visible image and highlight the effective information of image. Second, the enhanced visible and infrared images are decomposed by NSCT to obtain the low frequency and high frequency sub-bands. The improved rolling guide filter is used to enhance the edge and other details of the high frequency sub-band of infrared image. Third, mean gradient strategy and fuzzy logic strategy are used to fuse high and low frequency sub-bands, respectively. Experiments show that the proposed fusion method has better results than other state-of-the-art methods in terms of information entropy, standard deviation, and mutual information.