Image Fusion Based on NSCT and Minimum-Local Mean Gradient

To address the problems of low contrast, detail loss, and target blur in the fusion of traditional infrared and visible images, this study uses the idea of non-subsampled contourlet transform (NSCT) to improve the weight function and fusion rules and thus develops a new fusion algorithm to realize the effective fusion of infrared and visible images. First, NSCT is used to decompose infrared and visible images at multiple scales to obtain the corresponding low-and high-frequency coefficients. Then, the improved minimization and local mean gradient rules are used to fuse the low- and high-frequency coefficients, respectively, and thus to obtain the corresponding optimal fusion coefficient. The obtained fusion coefficient is then converted via an NSCT inverse transformation to obtain the final fused image. Finally, a public dataset is used to compare the proposed algorithm with the other five algorithms. The effectiveness and robustness of the proposed algorithm are verified under the constraints of seven performance evaluation indices having practical significance.