Research goal

• Improve drought injury detection in turfgrass using aerial multispectral imagery
• Can multispectral cameras detect drought injury that can not be seen in the visible spectrum?
• Can we detect drought injured turfgrass earlier before the effects are seen in the visible spectrum?

Personal involvement

• Develop pre-processing and machine learning pipeline for drought injury classification in turfgrass with multispectral images using Python and XGBoost
• Compare multispectral classification performance with RGB images to understand detection performance gain using all spectral bands

For a percent green calculator, I have written a script to process directories of lightbox images in Python for those who may not use ImageJ/Fiji regularly.