2025/02/26 by Romil Audhkhasi, Ningzhi Xie, Audhkhasi, Romil +5 · 3 citations
Computer Science · Engineering · Physics and Astronomy · #FOS: Physical sciences #Neural Networks and Reservoir Computing #Optical Polarization and Ellipsometry #Optics (physics.optics) #Photonic Crystals and Applications
paper · pdf · doi:10.48550/arxiv.2502.18739
openalex publication_date 2025/02/26 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Real time, singleshot multispectral imaging systems are crucial for environment monitoring and biomedical imaging. Most singleshot multispectral imagers rely on complex computational backends, which precludes real time operations. In this work, we leverage the spectral selectivity afforded by engineered photonic materials to perform bulk of the multispectral data extraction in the optical domain, thereby circumventing the need for heavy backend computation. We use our imager to extract multispectral data for two real world objects at 8 predefined spectral channels in the 400 to 900 nm wavelength range. For both objects, an RGB image constructed using extracted multispectral data shows good agreement with an image taken using a phone camera, thereby validating our imaging approach. We believe that the proposed system can provide new avenues for the development of highly compact and low latency multispectral imaging technologies.