Forage yield and quality estimation by means of UAV and hyperspectral imaging

Abstract This study investigated the potential of in-season airborne hyperspectral imaging for calibration robust forage yield and quality estimation models. An unmanned aerial vehicle (UAV) a imager were used to capture canopy reflections grass-legume mixture in range 450 nm 800 nm. Measurements performed over two years at locations Southeast Central Norway. All images subject radiometric geometric corrections before being processed ortho-images, carrying reflectance information. The data (n = 707) was split two, using half model remaining validation. Several powered partial least squares regression (PPLSR) models fitted estimate fresh (FM) dry matter (DM) yields, as well crude protein (CP), digestibility (DMD), neutral detergent fibre (NDF), indigestible (iNDF) content. Prediction performance these compared with prediction simple linear (SLR) models, which based on selected vegetation indices plant height. highest accuracies general pooled data, achieved by means PPLSR, relative root-mean-square errors validation 14.2% (2550 kg FM ha ?1 ), 15.2% (555 DM 11.7% (1.32 g CP 100 DM), 2.4% (1.71 DMD 4.8% (2.72 NDF 12.8% iNDF DM) FM, DM, CP, DMD, NDF, content, respectively. None tested SLR acceptable accuracies.