Reconfigurable Snapshot Hyperspectral Imaging Sensor\nBased on Monochromatic Pattern Match of Gradient Geometry Metasurface

Metasurface-based\nhyperspectral imaging has profoundly impacted\nthe field, with its easy integration and miniaturization advantages.\nNotably, metallic metasurfaces have been a research subject in sensing\nand imaging due to their strong capability of restraining light and\nsupporting high-fidelity and cost-effective manufacturing processes.\nIn this work, a novel metallic metasurface with gradient geometry\nwas proposed for hyperspectral imaging, which highlights the advantages\nof defect tolerance and reconfigurability of imaging units with a\nspectral image resolution of 135 × 75 superpixels. An orthogonal\nmatching algorithm with postcalibration based on monochromatic transmittance\nimages of the metasurface was designed for spectral reconstruction.\nThe proposed reconstruction algorithm simplifies calibration and can\ncompensate for systematic losses. The operating bandwidth of the hyperspectral\nimaging sensor can be reconfigured, and the miniaturized nature makes\nit potentially integrated into smartphones and other instruments with\nlightweight computing and fast reconstruction. Moreover, the strategy\nhas been validated in sensing and real-object surface imaging with\na spectral resolution of approximately 1.99 nm under a 2 nm line width.\nIts efficacy is verified in the 400–800 nm band with a high\nfidelity of up to 86.7%.