The Datacube Reconstruction Approach for Compressed Sensing Image Mapping Spectrometer (CSIMS)

Compressed sensing image mapping spectrometer can acquire the entire three-dimension (3-D) datacube of the objects instantaneously in a snapshot. The system can slice and encode the input image to different parts and disperses each part to be imaged on the detector. The reconstruction algorithm utilizes the sparse of the datacube in spatial and spectral dimensions to estimate the optimal results by compressed sensing concept. This paper establishes the sensing matrix for a novel datacube reconstruction approach, which combines the whole sliced parts to make full use of the raw mixture data. The simulations are conducted under different values of system parameters. The results show that the more pieces is sliced from the input image, the quality of the reconstructed datacube is higher, which reveals that the reconstruction approach proposed in this paper is effective to improve the quality of the datacube.