Abundance Estimation of Hyperspectral Data with Low Compressive Sampling Rate

Abstract

Hyperspectral data processing typically demands enormous computational resources in terms of storage, computation, and I/O throughputs. In this paper, a compressive sensing framework with low sampling rate is described for hyperspectral imagery. It is based on the widely used linear spectral mixture model. Abundance fractions can be calculated directly from compressively sensed data with no need to reconstruct original hyperspectral imagery. The proposed abundance estimation model is based on the sparsity of abundance fractions and an alternating direction method of multipliers is developed to solve this model. Experiments show that the proposed scheme has a high potential to unmix compressively sensed hyperspectral data with low sampling rate.