Weighted gradient-based fusion for multi-spectral image with steering kernel and structure tensor

In this paper, a novel weighted gradient-based fusion method combining structure tensor with steering kernel for multi-spectral images is proposed. The structure tensor is usually used to extract the main gradient information of a single point from the source image. However, the spatial structure information is important for human visual perception and the spatial context is an effective way of improving fusion result. Therefore we propose to use steering kernel to describe the spatial structure information of the source images, which is robust to the noise, and then use the weighted structure tensor that is combined with the steering kernel though a saliency metric in the proposed weighted gradient-based fusion method. Experimental results demonstrate that the proposed method outperforms other conventional fusion methods in term of visual comparison and quantitative assessment.