ON THE DERIVATION OF CROP HEIGHTS FROM MULTITEMPORAL UAV BASED IMAGERY
Keywords: RGB UAV imagery, 3D crop structure, point cloud, crop height raster, crop growth
Abstract. In this paper, we investigate the usage of unmanned aerial vehicles (UAV) to assess the crop geometry with special focus on the crop height extraction. Crop height is classified as a reliable trait in crop phenotyping and recognized as a good indicator for biomass, expected yield, lodging or crop stress. The current industrial standard for crop height measurement is a manual procedure using a ruler, but this method is considered as time consuming, labour intensive and subjective. This study investigates methods for reliable and rapid deriving of the crop height from high spatial, spectral and time resolution UAV data considering the influences of the reference surface and the selected crop height generation method to the final calculation. To do this, we performed UAV missions during two winter wheat growing seasons and generate point clouds from areal images using photogrammetric methods. For the accuracy assessment we compare UAV based crop height with ruler based crop height as current industrial standard and terrestrial laser scanner (TLS) based crop height as a reliable validation method. The high correlation between UAV based and ruler based crop height and especially the correlation with TLS data shows that the UAV based crop height extraction method can provide reliable winter wheat height information in a non-invasive and rapid way. Along with crop height as a single value per area of interest, 3D UAV crop data should provide some additional information like lodging area, which can also be of interest in the plant breeding community.