A Heterogeneous Image Fusion Method Based on Multi-directional Gradient Filtering

Abstract

Aiming at the problems of missing features, blurred edges, susceptible to noise interference and unsmooth grayscale transition in traditional heterogeneous image fusion methods, a fusion method of heterogeneous images based on multi-directional gradient filtering is proposed. The method uses the multi-scale of Laplacian pyramid and the multidirection of gradient filtering to extract the details of the source image to prevent the grayscale transition from being unsmooth; an adaptive fuzzy coupling mechanism is introduced to solve the problem that the source image features are easily missing; The gradient saliency mapping strategy avoids edge blurring and noise interference. The experimental results show that the method can achieve effective fusion of heterogeneous images.