SVH-BD : Synthetic Vegetation Hyperspectral Benchmark Dataset for Emulation of Remote Sensing Images

A synthetic hyperspectral dataset of 10,915 image cubes paired with pixel‑level vegetation biophysical parameters and uncertainty estimates. Generated through Sentinel‑2 preprocessing, PROSAIL inversion, and forward hyperspectral simulation, the dataset supports research on radiative transfer emulation, vegetation trait retrieval, uncertainty propagation, and spectral–biophysical modeling across four ecologically diverse regions.