PROJECT TITLE :

Denoising Of Hyper-spectral Image Using Low-Rank Matrix Factorization - 2017

ABSTRACT:

Restoration of hyperspectral pictures (HSIs) is a difficult task, thanks to the reason that images are inevitably contaminated by a mix of noise, as well as Gaussian noise, impulse noise, dead lines, and stripes, throughout their acquisition method. Recently, HSI denoising approaches primarily based on low-rank matrix approximation became a vigorous research field in remote sensing and have achieved state-of-the-art performance. These approaches, but, unavoidably need to calculate full or partial singular value decomposition of huge matrices, leading to the relatively high computational cost and limiting their flexibility. To address this issue, this letter proposes a methodology exploiting an occasional-rank matrix factorization theme, in that the associated strong principal element analysis is solved by the matrix factorization of the low-rank part. Our methodology desires solely an higher sure of the rank of the underlying low-rank matrix instead of the precise value. The experimental results on the simulated and real data sets demonstrate the performance of our technique by removing the mixed noise and recovering the severely contaminated pictures.