부산대학교 도서관

In this study we evaluated the potential of the hyperspectral sensor “Cubert UHD 185 Firefly” for yield prediction in 76 plots of a field trial with different varieties of winter barley at Königslutter (Lower-Saxony, Germany) in 2017. An UAV was used as carrier platform for the sensor. In 2017 we used 63 channels in a wavelength range of 450 to 700 nm. Predicted yield using PLSR and reference yield closely agreed with $\mathrm{R}^{2}=0.78$. We also calculated the NDVI RGB and evaluated its suitability for yield prediction in the same field trial. NDVI RGB and reference yield were less well related to each other with $\mathrm{R}^{2}=0.46$. The results show that using additional information from hyperspectral datasets allowed for a better yield prediction compared to RGB data alone. In 2018 a field trial with 76 plots of winter barley at Poppenburg (Lower-Saxony, Germany) was assessed on June 6, using the complete wavelength range of the sensor from 450 to 950 nm. In 2018, predicted yield using PLSR and reference yield agreed with $\mathrm{R}^{2}=0,81$.