Hyperspectral remote sensing image classification method

The invention relates to the field of remote sensing classification, and discloses a hyperspectral remote sensing image classification method, which comprises the following steps of: constructing an initialized SDAE model, deducing an objective function according to the number of iterations and a learning rate, judging a value of the learning rate according to a learning rate proportion, and judging whether learning can be ended or not according to a weight matrix and an offset. When the method provided by the invention is applied to remote sensing image classification, falling into a local optimal trap can be avoided, the condition of low manual operation efficiency is relieved, the classification precision is improved, and the experiment duration is shortened; the self-adaptive learning rate calculation method provided by the invention can help to learn image features deeper, improves classification precision, is high in learning efficiency, and can effectively solve the problems of local optimum, manual pa