C3SI-compact complementary compressed spectral imaging.

Abstract

Compressed spectral imaging (CSI) typically captures spectral images in a single shot by encoding scenes with a mask, but it often neglects information occluded by the mask. This paper capitalizes on such overlooked information to enhance spectral imaging performance. By integrating a sparse mask design, our approach captures both spectral data and complementary spatial information simultaneously. We utilize a spatial-spectral dual filtering technique to achieve superior reconstruction quality of spectral images. Contrasting with conventional dual-camera CSI systems, our design combines information encoding and beam splitting within the single device, circumventing the degradation of spectral signals typically caused by beam splitters. Through the integrated optical structure-mask-algorithm approach, it significantly improves reconstruction performance compared to existing CSI systems.