“In the field of hyperspectral unmixing, experts have proposed an unsupervised unmixing framework based on adversarial autoencoder networks, which effectively improves the unmixing accuracy and robustness.”