Hyperspectral image band selection method based on grouping sparsity

The invention discloses a hyperspectral image waveband selection method based on grouping sparsity. The method comprises the following steps: step 1, constructing grouping sparsity prior regularization; step 2, constructing an indication matrix with labels; and step 3, wave band selection of the hyperspectral image. The wave band selection method provided by the invention is a method capable of being used for carrying out wave band selection on important wave bands in the hyperspectral image, the wave bands which are seriously influenced by noise in the hyperspectral image can be effectively removed, and meanwhile, important information in the original hyperspectral image can be completely reserved. After waveband selection is carried out on the hyperspectral image by using the waveband selection method provided by the invention, the performance of a downstream classification task can be further improved. Therefore, the method has the characteristics of high practical application value and high reliability, a