Corn Emergence Uniformity at Different Planting Depths and Yield Estimation Using UAV Imagery

Uniform corn emergence is critical for maintaining optimum yield. Historically, studies relating plant emergence and early growth to yield for different treatments and management were based on small plots due to limitations in labor and time for in-field evaluation. Precision agriculture technologies (e.g., proximal sensors, yield monitors, and unmanned-aerial-vehicle (UAV)-based remote sensing) have now enabled field-scale evaluation. This study aimed to demonstrate UAV imagery applications in corn production at field scale with two case studies: 1) investigating corn emergence spatial variability at different planting depths; 2) estimating corn yield using image features. Red-green-blue and multispectral images were acquired to determine emergence parameters and vegetation indices (VIs) for early plant growth (V4, V6, and V7) indicators. In case study 1, average and coefficient of variation of emergence parameters were computed in 1.0 m x 6.1 m grids for four planting depths. In case study 2, yield was estimated by different feature datasets in random forest models. Results demonstrated that there was spatial variability within the planting depth treatments along the transects. Emergence data alone could not explain variation in yield (R2 = 0.01); however, the combination of VIs at all growth stages could estimate yield with R2 of 0.34. These case studies demonstrated UAV imagery usage in studying crop emergence variability and estimating yield at field scale. Future studies should include more timely UAV data along the growing season in different fields and years to develop a more robust model.