Quantification of species composition in grass-clover swards using RGB and multispectral UAV imagery and machine learning

• DOI: 10.3389/fpls.2024.1414181
• Publication Date: 2024-06-19T04:40:28Z
• Authors: Joanna Pranga; Irene Borra-Serrano; Paul Quataert; Tom De Swaef; Thijs Vanden Nest; Koen Willekens; Greet Ruysschaert; Ivan A. Janssens; Isabel Roldán-Ruiz; Peter Lootens
• Keywords: RGB color model; Red Clover

Introduction Growing grass-legume mixtures for forage production improves both yield productivity and nutritional quality, while also benefiting the environment by promoting species biodiversity enhancing soil fertility (through nitrogen fixation). Consequently, assessing legume proportions in mixed swards is essential breeding cultivation. This study introduces an approach automated classification mapping of grass-clover using object-based image analysis (OBIA). Methods The OBIA procedure was established RGB ten band multispectral (MS) images capturedby unmanned aerial vehicle (UAV). workflow integrated structural (canopy heights) spectral variables (bands, vegetation indices) along with a machine learning algorithm (Random Forest) to perform segmentation classification. Spatial k-fold cross-validation employed assess accuracy. Results discussion demonstrated good performance, achieving overall accuracy approximately 70%, MS-based imagery, grass clover classes yielding similar F1 scores, exceeding 0.7 values. effectiveness examined analyzing correlations between predicted fractions dry matter (DMY) proportions. quantification revealed positive strong relationship, R2 values 0.8 outcomes. indicates potential estimating (relative) coverage, which could assist breeders but farmers precision agriculture context.