Hyperspectral Unmixing Via Wavelet Based Autoencoder Network

Hyperspectral unmixing is a hot topic in the field of remote sensing. Due to the limitation of spatial resolution and diversity of object distribution, hyperspectral image contains mixed pixels, which brings a great challenge for hyperspectral image processing. A novel wavelet based hyperspectral unmixing autoencoder network is proposed in this paper. In the framework of autoencoder network, multiscale wavelet coefficients of the signal are employed, which contribute to learn the intrinsic feature of endmember deeply. Moreover, the cost function of the network is designed according to the sparsity and nonnegative constraints of abundance, as well as the spectral fidelity. Experimental results on both simulated and real hyperspectral data sets demonstrate that the proposed method outperforms other state-of-the-art unmixing methods.