Adaptive weighted image fusion algorithm based on NSCT multi-scale decomposition

The purpose of infrared image and visible light image fusion is to preserve as much as possible the target information in the infrared image and the detailed information in the visible light image. For this purpose, this paper first extracts the infrared image saliency map based on guided filtering, uses NSCT to decompose infrared and visible light images at multiple scales, adds the infrared image saliency map to the low-frequency component, and applies weighted average fusion through the adaptive infrared weight map. In the high frequency component, the absolute value of the maximum coefficient is taken to be the maximum for fusion. This method takes into account the generally high brightness of human targets in infrared images, maintains the edge information of the image, and fully considers the anti-noise performance of the fusion method, and can adaptively take into account the complex and changeable environment.