Abstract:Aiming at the classification problem of hyperspectral data in remote sensing images, a hyperspectral remote sensing image classification method based on the deep learning feature representation by stacked sparse auto-encoder (SSAE) is proposed. First, the spectral data samples are pre-processed and normalized. Then, it is input into the SSAE for feature representation learning, and the grid search is used to obtain the optimal network parameters, thereby obtaining a valid feature representation. Finally, the input image features are classified by the support vector machine (SVM) classifier, and finally the pixels in the remote sensing image are classified. Experimental results on two standard datasets show that this method can achieve accurate hyperspectral landmark classification.