Hyperspectral image noise reduction method based on low-rank tensor two-factor joint optimization

The invention discloses a hyperspectral image denoising method based on low-rank tensor two-factor joint optimization, which comprises the following steps of: establishing a denoising optimization model based on the combination of image prior and a factorization method according to a generalized denoising model definition formula of a hyperspectral image, and introducing L2, L2,..., L2; the p norm is used as a constrained regular term, and a partition approximation algorithm is used to convert the optimization model into a form of alternate iteration updating of a plurality of sub-problems; then, respectively carrying out optimization solution on each sub-problem, respectively solving each sub-problem by using tube fiber decomposition, a Silverst matrix equation, a singular value decomposition method, a fast Fourier transform method and the like, and finally, outputting a de-noising result by checking whether a convergence condition is met or not. Through the method, the original pure image can be restored un