Using RGB remote sensing for biomass prediction in temperate grassland

摘要

Grassland monitoring techniques which can predict biomass yield over the whole farm enable the farmer to make site-specific management decisions. Crop biomass and plant height are highly correlated with each other and thus, crop surface height can beused for yield prediction. Non-destructive remote sensing tools are useful for obtaining spatial information on plant height on larger scales. In this study a red, green, blue (RGB) camera on a small unmanned aerial vehicle (UAV) was used for yield prediction ina temperate grassland experiment. The experiment was established in 2016 with two legume-grass mixtures (clover-grass, lucerne-grass) as well as in pure stands of legumes and grasses. RGB imaging was carried out every second week for estimating the growth rate of the grasslands and before each harvest (4 cuts). In parallel, reference data for biomass yield and manual height measurements were taken. 3D crop surface models generated from RGB imaging based measurements were used for calculating thecrop surface height. The biomass yield was predicted with satisfactory results.