Work

High-definition (HD) snapshot diffractive computational spectral imaging and inferencing

Year: 2025

Type: article

Abstract: Capturing spatial and spectral information from a scene is challenging due to the seemingly orthogonal nature of these dimensions. Here, we introduce a compact high-definition snapshot diffractive spectral imaging system that employs a diffractive filter array to project spectral information into the spatial domain, enabling the capture of a single-channel 2D image that contains both spatial and spectral data. This single diffractogram undergoes computational reconstruction into a spectral image cube, providing a high-resolution depiction of the scene across 25 spectral channels within the 440–800 nm range, containing 1304×744pixels (∼1MP). Importantly, in our current implementation, the camera’s field-of-view and spatial resolution are primarily determined by the conventional optics and sensor, achieving ∼50 ∘ and 2.9µm, respectively. Our approach offers several advantages: instantaneous snapshot capture, optical compression, flexible offline reconstruction, post-capture spectral basis selection, ultra-compact implementation, inferencing without spectral reconstructions, and high light throughput due to the absence of lossy filters. We achieve a spectral reproduction error of less than 15% across the 440–800 nm band. Proof-of-concept experiments across diverse applications—such as biological tissue classification, food quality assessment, and simulated stellar photometry—demonstrate our system’s inference capabilities. These findings position diffractive computational spectral imaging as a versatile modality for advancing scientific and industrial imaging applications.

Source: Optica

DOI: 10.1364/optica.559279