A Fast Snapshot Hyperspectral Image Reconstruction Method Based on Three-Dimensional Low Rank Constraint

摘要

The snapshot hyperspectral imaging is an emerging technique with numerous applications.However, the hyperspectral imaging reconstruction is often time-consuming, which is placinga limit on the development of snapshot hyperspectral imaging. We present an efficientreconstruction algorithm based on the tensor analysis and the low-rank constraint. Thehyperspectral data cube is regarded as a low rank three-order tensor, which can jointly treatboth spatial and spectral modes. The 3D-LRC method can greatly decrease the computationtime without unfolding the hyperspectral data cube into 2D patches. Compared withthe-state-of-the-art method, the proposed method has a great improvement in the reconstructionspeed and quality. The method has been implemented on two typical snapshothyperspectral imaging systems.