A VIS-NIR multispectral endoscopic imaging system based on computational imaging and parallel acquisition

Abstract

Multispectral endoscopic imaging enables simultaneous acquisition of spatial and spectral information for functional tissue assessment. Conventional multispectral endoscopes are often limited by bulky optical architectures and restricted spectral channels. Here, we demonstrate an active single-pixel multispectral endoscopic system based on parallel spectral detection. Structured illumination generated by a digital micromirror device is efficiently coupled into the endoscopic channel, while multiple narrowband-filtered detectors integrated at the probe tip enable synchronous visible–near-infrared (VIS-NIR) acquisition. Fourier single-pixel reconstruction is employed to achieve high-quality imaging at low sampling rates. Experiments under white-light, broadband, and fluorescence conditions confirm stable parallel acquisition, minimal crosstalk, and reliable spectral discrimination from 450 to 1550 nm. The proposed architecture provides a compact and scalable approach for VIS-NIR multispectral endoscopy.