Funded by the Illinois Soybean Association checkoff program.

STUDENT RESEARCHER

STUDENT RESEARCHER

Xian Liu

M.S. Level Student
Southern Illinois University Carbondale
xianliu2014@siu.edu
Advised by Dr. Jason Bond

Using Multiple Remote Sensing Platforms To Monitor and Manage Soybean Cyst Nematode Disease

As one of the most damaging soybean pathogens in the United States, the soybean cyst nematode (SCN) causes over 3 billion dollars annual losses. The ever-adapting nature of SCN populations makes it more challenging on disease management with available options for this pathogen such as using soybean varieties resistant. Recent developments and advances in multispectral imaging technology make remote sensing a promising tool to traditional crop disease detection and monitoring. This study aimed to develop an early detection methodology with machine learning to monitor and assess the impact of SCN on crop health. Field trails were established at SCN infested farms at multiple locations in Southern Illinois to collect SCN egg counts and multispectral imagery of SCN-related crop stress. Multispectral images were collected with DJI Matrice 210 drone mounted with an Altum multispectral camera at field level on a weekly basis. Soil samples were collected at planting, midseason, and harvest to determine the SCN population densities. Yield data were collected at the end of the season at soybean harvest. Microplots were set at university research center to collect hyperspectral data with ASD hand-held spectroradiometer at microplot level of varying SCN infections. A total of 16 common UAV-based Vegetation Indices (VI) was selected to examine their relationships with SCN infestation and yields. Enhanced Vegetation Index (EVI) and Difference Vegetation Index (DVI) were reported to have the highest correlations with both SCN egg counts and crop yields according to the Pearson correlations. We found that red and near infrared bands were most sensitive to SCN infections. An unsupervised machine learning and supervised deep learning approach will be utilized to identify SCN-infested areas. With studying the drone images, hyperspectral reflectance, and ground truth data, this research will develop a set of algorithms to assess the impact of SCN on crop health.