Spectral-spatial classification of hyperspectral imagery based on deep convolutional network

Hyperspectral image (HSI) classification has been an active topic in recent years. Over the past few decades, a significant number of methods have been proposed to deal with this problem. However amongst these methods, deep learning based methods are rare. Inspired by the excellent performance of deep convolutional neural network (DCNN) in visual image classification, in this paper, we introduce DCNN into HSI classification. Instead of using two-dimension kernels as DCNN is used in two-dimension image classification, one-dimension kernels is adopted in our DCNN to fit the HSI context. The proposed method is compared with the state-of-the-art deep learning based HSI classification methods, evaluated on two popular datasets, and produces better classification results.