Early detection of Soybean Sudden Death Syndrome through Satellite Images

Abstract

Sudden death syndrome (SDS) is one of the major yield-limiting soybean diseases in the Midwest. Effective management for SDS requires accurate detection in soybean fields. Since traditional scouting methods are time-consuming, labor-intensive and often destructive, alternative methods to monitor SDS in large soybean fields are needed. This study explores the potential of high-resolution (3 m) PlanetScope satellite imagery for early and accurate detection of SDS using a random forest classification algorithm. We used four spectral bands including red, blue, green, and near-infrared (NIR) and calculated normalized difference vegetation index (NDVI) to detect healthy and SDS-infected quadrats (3 m wide by 1.5 m in length) in a soybean field experiment located in Boone, Iowa. Data collected during the 2016, 2017 and 2018 soybean growing seasons were analyzed in this study. The results indicate that spectral bands of high-resolution PlanetScope imagery along with calculated NDVI can accurately predict SDS in soybean plots even before foliar symptoms develop. Healthy and diseased soybean quadrats were detected with more than 85% accuracy and with kappa statistics, a measure of inter-rater agreement, of more than 68% in all growing seasons. These promising results suggest that high-resolution satellite imagery has tremendous potential for detection of SDS in soybean fields. Our findings highlight that this technology can facilitate large-scale monitoring of SDS and possibly other economically important soybean diseases to guide recommendations for site-specific management in current and future seasons.