Higher Speed Multispectral Imaging Using a Thin Liquid Crystal Layer in Metal Cavity and Artificial Intelligence Reconstruction Algorithm

Abstract A concept and demonstration of a fast multispectral imager operating across the visible (Vis) and near‐infrared (NIR) ranges is presented. This system utilizes a specially designed LC tunable metal‐insulator‐metal (MIM) cavity combined with a passive multiband pass filter (MBF). The integration of the MBF with the MIM device, along with advanced artificial intelligence algorithms, provides several key benefits: i) lower production cost, ii) dual functionality of the metal layers as reflectors and electrodes, iii) a richer spectrum exhibiting a larger number of peaks and valleys at reduced thickness, enabling faster switching and robust spectrum reconstruction, iv) a broader operational spectrum, v) higher light throughput as only one polarizer is needed, vi) reduced tolerance requirements on the contrast of the spectrum due to the use of artificial neural nets reconstruction algorithm, and vii) lower operation voltage. The MIM tunable device exhibits as fast as 50 µs stabilization time with an average value of 250 µs, and a total acquisition time of <5 ms, highlighting the potential of this technology for high‐rate multispectral video. The system's applicability is demonstrated in food monitoring, showcasing its potential for real‐time efficient spectral analysis.