Remote thermal infrared imaging for rapid screening of sudden death syndrome in soybean

Abstract

Sudden death syndrome (SDS), a fungal infection in soybean caused by Fusarium virguliforme, greatly affects the plant health and in some cases, can cause yield losses of more than 70%. Infected plants are scored by visual assessment based on severity and extent of infection. This manual process is time intensive and not practical for large acreages. Diseased and stress in plants show elevated canopy temperatures that can potentially lead to identification of unhealthy plants without manual scoring. The infection decreases nutrient distribution causing stress that results in internal plant temperature to increase. Thermal infrared (TIR) sensors have the ability to measure the emitted radiation of an object in the infrared region of the electromagnetic spectrum to estimate canopy temperatures. However, TIR sensors have not yet been utilized to capture changes in canopy temperatures to detect SDS in soybean. Therefore, the goal of this study was to (1) use a TIR sensor to assess plant health and vitality, and (2) evaluate canopy temperatures over the growing season to quantify disease development. A thermal infrared camera was mounted on a small unmanned aerial system to capture aerial imagery over the growing season. The first flight was achieved once SDS foliar symptoms began initial development. The remaining three flights occurred before, during, and after full pod fill when symptoms had reached their apex. Results show increasing correlations over the four days. Elevated canopy temperature changes were observed on canopies at early SDS symptom development. Symptoms at the end of the growing season displayed strong correlations to the canopy temperature with ρ = 0.7404. Disease severity showed the strongest correlation throughout the four flights with the last at ρ = 0.7245. The four flights exhibit a decreasing trend with Spearman's rho (R2 = 0.86 for disease severity). Therefore, thermal imaging can be utilized to detect diseased plots. Future studies will be conducted to understand how to mitigate for SDS using thermal detection.