PRISM: A Polyhedral Reflective Imaging Scanning Mechanism for Hyperspectral Applications

Hyperspectral imaging has become essential in fields such as waste sorting and agriculture due to its ability to capture detailed spectral information. However, conventional hyperspectral imaging systems typically rely on mechanical movements, such as conveyor belts or linear stages, to scan the target objects, which limits their applicability in environments with static targets or space constraints. To address these challenges, we present PRISM (Polyhedral Reflective Imaging Scanning Mechanism), a novel system that facilitates rapid and high-precision imaging of stationary objects by leveraging a polyhedral reflective structure, thereby eliminating the need for external relative motion. Experimental evaluations demonstrate that PRISM can achieve full spatial-spectral data of 512 by 727 pixels with 448 bands in 2.01 seconds, while maintaining high spectral accuracy, with a reflectance standard deviation of 0.015. We further validate PRISM's effectiveness through textile waste sorting experiments, where PRISM accurately differentiates between various material types at both the pixel and object levels, showcasing its potential for practical applications in industrial settings.