Satellite hyperspectral image denoising method, device and equipment based on self-supervised learning

The invention provides a satellite hyperspectral image denoising method, a satellite hyperspectral image denoising device and satellite hyperspectral image denoising equipment based on self-supervised learning. The satellite hyperspectral image denoising method comprises the following steps: acquiring a satellite hyperspectral image to be processed; sampling the to-be-processed satellite hyperspectral image to obtain at least two first sampling instances; and based on the at least two first sampling instances, adopting a pre-training denoising model to obtain a denoising result of the to-be-processed satellite hyperspectral image. The denoising model used in the invention is obtained through training based on the self-supervised learning algorithm, so that the training of the denoising model can be completed under the condition that a clean/noise image pair of a real satellite hyperspectral image cannot be obtained, and then the denoising processing of the real satellite hyperspectral image is realized throug