A dataset of hyperspectral Look-Up Tables for 3.5 million traits and structural combinations of Central European temperate broadleaf forests
| Authors | |
|---|---|
| Year of publication | 2024 |
| Citation | |
| Description | The dataset comprises 3.5 million unique combinations of leaf biochemical properties and canopy structural characteristics of forest scenes, along with various configurations of solar geometry. It provides spectral data spanning wavelengths from 450 nm to 2300 nm, with a 2 nm resolution. The dataset is divided into two files: one detailing the average reflectance of all scene pixels and another focusing specifically on sunlit leaf pixels. Look-up tables (LUT) were created using version 5.10.0 of the Discrete Anisotropic Radiative Transfer model. The virtual forest scenes are based on 3D models of European beech trees, derived from Terrestrial Lidar Scanning, and adjusted for different leaf area indices and structural configurations to simulate natural forest diversity. Reflectance data were processed using Matlab and Python scripts, resulting in hyperspectral cubes, which were subsequently used to generate the LUT. This dataset is valuable for training machine learning models aimed at retrieving forest functional traits and for calibrating remote sensing algorithms. Its high spectral resolution and diverse trait combinations make it adaptable to different times, locations, and both hyper- and multispectral sensors, supporting future hyperspectral satellite missions. |
| Related projects: |