Find in Library

The goal of multispectral imaging is to obtain the spectrum for each pixel in the image of a scene and deliver much reliable information. It has been widely applied to several fields including mineralogy, oceanography and astronomy. However, multispectral images (MSIs) are often corrupted by various noises. In this paper, we propose a MSI denoising model based on nonlocal multitask sparse learning. The nonlocal self-similarity across space and the high correlation of the MSI along the spectrum via multitask sparse learning are fully exploited in the proposed model. A nonnegative matrix factorization (NMF) based algorithm is developed to solve the proposed model. Experimental results on both simulated and real data demonstrate that the proposed method performs better than several existing state-of-the-art denoising methods.