Principal Component Extraction for Remote Sensing Image Using Auto-Association Neural Network
自联想神经网络的遥感图象主分量提取

A new method of principal component extraction for remote sensing image based on auto-association neural network is presented in this paper. The proposed method is applied to the image data compression, feature extraction and filtering. Experiments show that auto-association neural network for extracting the principal components of the image is as good as the conventional K-L transform.