High-throughput plant height measurement for the field peanuts from low-cost UAV photogrammetry

Abstract. Plant height is, as a crucial indicator, capable of reflecting the health status and growth vigor at various growth stages. It provides essential information for increasing crop yield, optimizing cultivation strategies, and improving varieties. Traditional plant height measurements using tapes or rods are labour-intensive, time-consuming, subject to human errors, and inadequate for large-scale observations. In recent years, unmanned aerial vehicles (UAVs) equipped with RGB cameras have demonstrated significant advantages in terms of efficiency and cost-effectiveness, enabling detailed 3D reconstruction of complex farmland environments through photogrammetry techniques. Therefore, we develop a high-throughput plant height measurement approach for the field peanuts from low-cost UAV photogrammetry. First, a UAV platform equipped with RGB camera is used to collect high-resolution imagery, covering the entire peanut growth stages. Following this, the aerial images are processed and precisely aligned with positional and orientation system (POS) data, subsequently generating Digital Surface Models (DSMs). Among these DSMs, the one representing the bare soil period was considered as the Digital Elevation Model (DEM). Afterwards, each plot is clipped based on its minimum bounding rectangles, creating Canopy Height Models (CHMs) by subtracting the DEM from the corresponding DSMs. Finally, Peanut plant heights are estimated via histogram distribution analysis of CHMs and validated with manually measured heights in Wangbian Community, Ningyang County, Tai'an City, Shandong Province. Experimental results indicate excellent effectiveness and reliability, achieving coefficients of determination (R2) of 0.9424 and RMSE of 2.26 cm. These observations demonstrate UAV photogrammetry's practical potential for large-scale crop phenotyping applications.