Hyperspectral mixed noise removal method based on dual total variation

The invention provides a hyperspectral mixed noise removal method based on double total variation, which comprises the following steps: acquiring a hyperspectral image containing mixed noise, selecting an SSTV regular term to describe a global sparse structure of the hyperspectral image, designing a reweighted norm to constrain a local sparse structure of a spatial direction difference image of a principal component diagram of the hyperspectral image at the same time, introducing double constraints, selecting sparse global features and local features, selecting sparse regular terms for isolating sparse noise, describing Gaussian noise of an image through norms, constructing a hyperspectral image denoising model based on double total variation, and introducing auxiliary variables to optimize the hyperspectral image denoising model based on double total variation. An ALM algorithm framework is adopted to optimize the proposed model, and a complex problem is converted into a plurality of simple sub-problems for