Improving the Performance of Compressive Spectral Imaging with Bayer Color Filter Array

Reconstructing hyperspectral images (HSIs) from coded measurements from coded aperture snapshot spectral imaging (CASSI) system is essential for capturing images that offer superior spectral resolution over traditional RGB images. Nevertheless, the reconstruction of HSIs is difficult because of the aliasing of spatial and spectral information in coded measurements. To address this issue, we explore the possibility of utilizing an RGB camera with a Bayer color filter array (CFA) as an alternative to the gray-scale camera for implementing CASSI, which we term Bayer-CASSI. We first formulate the mathematical model of Bayer-CASSI, and then we evaluate the performance of Bayer-CASSI in simulation data and real data captured by our self-built Bayer-CASSI optical system. Experiment results in both simulation and real data show the effectiveness of Bayer-CASSI.