Paired VNIR hyperspectral and RGB images of canadian wild rice kernels

This dataset contains paired VNIR hyperspectral data and corresponding RGB images of individual Canadian wild rice (Zizania palustris) kernels across eight quality categories. VNIR data were acquired using a Specim FX10e scanner (400–1000 nm) and processed to reflectance-corrected hypercubes with per-kernel segmentation masks and mean spectral profiles. The paired RGB subset includes one RGB image per kernel, aligned one-to-one with VNIR samples using consistent indexing. The dataset supports multimodal analysis, kernel-level spectroscopy, and machine learning research in agricultural quality assessment.