Related Experiment Video
Updated: Aug 1, 2025
11:37
RGB and Spectral Root Imaging for Plant Phenotyping and Physiological Research: Experimental Setup and Imaging Protocols
Published on: August 8, 2017
16.3K
Evaluation of Stem Rust Disease in Wheat Fields by Drone Hyperspectral Imaging
Jaafar Abdulridha1, An Min1, Matthew N Rouse2
1Bioproducts and Biosystems Engineering Department, University of Minnesota, 1390 Eckles Ave, St. Paul, MN 55108, USA.
Sensors (Basel, Switzerland)
|April 28, 2023
Summary
Detecting wheat stem rust early using drone hyperspectral imaging is crucial for crop management. This technology accurately identifies disease severity, aiding breeders and farmers in protecting cereal crops.
Area of Science:
- Agricultural Science
- Plant Pathology
- Remote Sensing Technology
Background:
- Cereal crops are vital for global food security, facing threats from diseases like wheat stem rust.
- Advanced technologies are needed for early and accurate disease detection to minimize chemical use and costs.
- Unmanned Aerial Vehicle (UAV) based hyperspectral imaging offers a promising solution for crop monitoring.
Purpose of the Study:
- To evaluate the effectiveness of drone-mounted hyperspectral imaging for detecting and classifying wheat stem rust severity.
- To identify optimal wavelengths and spectral vegetation indices (SVIs) for disease detection.
- To assess the potential for developing cost-effective sensors for disease diagnosis.
Main Methods:
- Utilized a hyperspectral camera on a UAV to survey 960 wheat plots with varying stem rust severities.
- Applied machine learning classifiers including Quadratic Discriminant Analysis (QDA), Random Forest Classifier (RFC), Decision Tree, and Support Vector Machine (SVM).
- Classified disease severity into four levels: healthy, mild, moderate, and severe, and performed binary classification for mild vs. non-diseased plots.
Main Results:
- Random Forest Classifier (RFC) achieved the highest overall classification accuracy of 85% for disease severity levels.
- RFC also yielded the best classification rate (76%) for spectral vegetation indices (SVIs), identifying Green NDVI, PRI, RVS1, and Chlorophyll Green.
- Binary classification of mild vs. non-diseased plots reached 88% accuracy, demonstrating sensitivity to low disease levels.
Conclusions:
- Drone-based hyperspectral imaging is effective in discriminating wheat stem rust severity levels.
- This technology enables efficient selection of disease-resistant wheat varieties by breeders.
- Early detection of wheat stem rust outbreaks by farmers is facilitated, allowing for timely management decisions.
