Spatial-Spectral Multiscale Sparse Unmixing for Hyperspectral Images

We propose a simple yet efficient sparse unmixing method for hyperspectral images. It exploits the spatial and spectral properties of hyperspectral images by designing a new regularization informed by multiscale analysis. The proposed approach consists of two steps. First, a sparse unmixing is conducted on a coarse hyperspectral image resulting from a spatial smoothing of the original data. The estimated coarse abundance map is subsequently used to design two weighting terms summarizing the spatial and spectral properties of the image. They are combined to define a sparse regularization embedded into a unmixing problem associated with the original hyperspectral image at full resolution. The performance of the proposed method is assessed with numerous experiments conducted on synthetic and real datasets. It is shown to compete favorably with state-of-the-art methods from the literature with lower computational complexity.

Automated summary

In this paper, we propose a simple yet efficient sparse unmixing method for hyperspectral images, which exploits the spatial and spectral properties of the images. The proposed method performs a sparse unmixing on a coarse hyperspectral image obtained through spatial smoothing, and then uses the estimated coarse abundance map to design a sparse regularization for the original hyperspectral image. Experimental results demonstrate that the proposed method achieves competitive performance with lower computational complexity compared to state-of-the-art methods.