HyperFusion: Full-Spectrum Multi-Modal Hyperspectral Imaging System for Food and Agricultural Intelligence

Moonshot Award

Proposal Development

Abstract

Advancing sustainable food and agricultural systems increasingly depends on data-driven, non-destructive sensing technologies capable of rapidly assessing the quality and physiological properties of biological materials. Hyperspectral imaging (HSI) provides rich spectral–spatial information related to biochemical composition; however, most HSI systems are restricted to either the visible–near-infrared (VNIR) or short-wave infrared (SWIR) region and operate using a single optical mode, limiting access to complementary surface and internal information. 

We developed HyperFusion, a full-spectrum, multi-modal HSI platform integrating VNIR and SWIR imaging with modular illumination and both reflectance and transmittance acquisition. Spectral and spatial calibration protocols were established to enable accurate imaging, spectral continuity, and cross-modal alignment. The platform was further integrated with automated acquisition and processing software for streamlined data collection, spectral correction, and multimodal fusion. 

As a case study, individual Concord grape berries across different maturity stages were analyzed using multimodal hyperspectral data. Soluble solids content (°Brix) and pH were predicted to evaluate the complementary value of spectral ranges and imaging modalities for non-destructive quality assessment. Overall, HyperFusion provides a flexible platform for investigating multimodal optical signatures and developing intelligent sensing approaches for food and agricultural applications.