为了有效利用高光谱遥感图像中的波段信息，提高高光谱遥感图像分类的精确度，本文提出了基于栈式自编码神经网络（ Stacked Autoencoder，SA）对高光谱遥感图像进行分类。栈式自编码神经网络充分利用高光谱图像中的光谱信息，对其进行相应特征提取，避免了相邻信息间的相关性和信息的冗余，本方法采用无监督学习和监督学习相结合，既可以像传统方法那样进行降维，简化相关的计算复杂度，同时在分类精度上有很大地提高。

In this study, we aim to utilize the band information in hyperspectral remote sensing images effectively and improve the accuracy of hyperspectral remote sensing image classification. The proposed method is aimed at the classification of hyperspectral remote sensing images based on a stacked autoencoder. A stack self-encoding neural network takes advantage of the spectral information in a hyperspectral image, extracts the corresponding features, and reduces the relativity between adjacent information and the redundancy. This method combines unsupervised learning with supervised learning. Hence, we can reduce the dimension and simplify the computation complexity as in traditional methods. The proposed algorithm improved the classification accuracy effectively.