Answers are grounded in this paper and cite their sources.
Compact metasurface hyperspectral image sensors have attracted great interest in recent years owing to their potential for high temporal-resolution sensing. However, conventional metasurface-based hyperspectral image sensors suffer from a fundamental trade-off between spectral and spatial resolutions due to the collective resonance effect that necessitates large periodic structures to achieve a narrow modulation function. To circumvent this limitation, we report a computational hyperspectral sensor based on longitudinally-coupled multilayer structures (LCM-HS imager) that simultaneously achieves high spatial and spectral resolutions. By optimizing inter-cavity coupling within a multilayer geometry, we demonstrate an ultra-compact (5.8 µm) spectral encoding chip, whose transmission linewidth is narrowed to 3 nm ( Q factor ∼175), with a spectral correlation width of ∼5nm. Through an upsampling strategy with spatial oversampling, the LCM-HS imager achieves an effective pixel pitch of 2.9 µm, allowing it to resolve finer structural details than commercial hyperspectral cameras. The negligible absorption of all-dielectric materials enables the spectral encoding chip to achieve a modulation depth of up to 90%, resulting in excellent reconstruction performance in the blue and green spectral bands under outdoor conditions. Driven by deep learning-based reconstruction algorithms, the LCM-HS imager achieves high-speed hyperspectral imaging at 28 fps, delivering a spatial resolution of 640×640 pixels across 41 spectral channels. Our sensor delivers exceptional spatial and spectral resolution, offering significant potential for snapshot spectral imaging in demanding applications such as biological fluorescence microscopy and Raman imaging.
No takes yet. Share an insight, caveat, or question.
Zhang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: