Tar Spot Disease Quantification Using Unmanned Aircraft Systems (UAS) Data

Tar spot is a foliar disease of corn characterized by fungal fruiting bodies that resemble tar spots. The emerged in the U.S. 2015, and severe outbreaks 2018 caused an economic impact on yields throughout Midwest. Adequate epidemiological surveillance quantification are necessary to develop immediate long-term management strategies. This study presents measurement framework evaluates severity using unmanned aircraft systems (UAS)-based plant phenotyping regression techniques. UAS-based phenotypic information, such as canopy cover, volume, vegetation indices, were used explanatory variables. Visual estimations performed expert pathologists per experiment plot basis response Three methods, namely ordinary least squares (OLS), support vector (SVR), multilayer perceptron (MLP), determine optimal method for measurement. cross-validation results showed model based MLP provides highest accuracy measurements. By training testing with spatially separated datasets, proposed achieved Lin’s concordance correlation coefficient (ρc) 0.82 root mean square error (RMSE) 6.42. demonstrated we could use spot, which shows gradual spectral develops.