A New Variational Model in Texture Space for Pansharpening

In this letter, a new variational model in texture space for pansharpening is proposed to increase spatial information of multispectral image, while preserving spectral and spatial consistencies of pansharpened image. Geometric structure consistency between panchromatic image and pansharpened image is very important for preserving spatial characteristics, especially in borders, edges, and textured regions. G-norm can extract more pure and accurate texture, curvature, and oscillating details than do previous first-order and second-order-based methods, such as Gradient and Hessian operators. Therefore, we aim to use the G-space to maintain spatial information. Experimental results show that the proposed method has a better performance in terms of both spatial and spectral qualities; however, it is not efficient in terms of computation time.