Forecasting Corn Yield Using Structure from Motion and Multispectral Data Via UAS.

Recent incursion of small-unmanned aerial systems (sUAS) offer high revisiting time and ultra-high resolution data filling a gap of information for in-seasonal m Structure from Motion (SfM), multispectral indices and a method for plant detection were evaluated as early and late season proxies (UAS metrics) for yield forecasting. An experiment was implemented at Ashland Bottom Experimental Station, Kansas on 2016. The study includes Nitrogen fertilization, population rates and different hybrids. All were utilized as base line for spatial and temporal variability between the UAS metrics and the yield response. A stepwise procedure was implemented for variable selection. Finally, a multivariate regression was implemented using UAS metrics at v4,v8 and vt stages and yield response to evaluate the potential of yield forecasting using this approach.