Full-Spectrum UAV Hyperspectral Imagery (VNIR-SWIR) for Wheat Phenotyping and Breeding Research
Published by CSIRO Data Access Portal
About this dataset
This collection contains full-spectrum UAV hyperspectral imagery acquired over wheat field experiments at 2024 BARS, 2025 BARS, and 2024 Kabinga using a Mjolnir VS-620 hyperspectral imaging system (HySpex) mounted on a BlackSquare Hercules X8 unmanned aerial vehicle (UAV). The system integrates co-mounted visible and near infrared (VNIR) and shortwave infrared (SWIR) imaging spectrometers, enabling collection of contiguous hyperspectral imagery across a broad spectral range suitable for crop phenotyping, physiological trait assessment, and agricultural research. Data acquisition was managed through the HySpex AIR interface, which is used to configure sensor acquisition parameters and flight operations. During each flight, imagery was collected alongside navigation and orientation data recorded by an APX-15 GNSS/INS system. The resulting navigation products are provided as T04 files, which can be processed using POSPac software to support geolocation, georectification, and orthorectification workflows. The collection includes raw hyperspectral image data, flight logs, navigation data, radiometrically calibrated radiance products, and a range of derived analytical outputs. Raw imagery is stored in proprietary HySpex formats with accompanying header files describing acquisition parameters. Radiance products were generated from raw digital number (DN) imagery using HySpex RAD software and manufacturer-provided calibration information. Subsequent processing workflows include VNIR-SWIR co-registration, reflectance derivation, plot-level extraction, spectral jump correction, and generation of representative canopy spectra. The collection contains both original and processed datasets to support data reuse, methodological development, independent validation, and reprocessing using alternative workflows. The imagery captures plot-scale spatial and spectral variation across wheat breeding and research experiments and may be used for applications including canopy reflectance characterisation, physiological trait prediction, crop phenotyping, machine learning model development, spectral feature analysis, and UAV hyperspectral sensing research. Collection Purpose This dataset was collected to support the development and evaluation of UAV hyperspectral sensing approaches for wheat breeding and crop phenotyping. The full-spectrum VNIR-SWIR imagery provides detailed spectral information for characterising crop canopies and investigating relationships between canopy reflectance, physiological traits, crop development, and agronomic performance.