Efficient Drone-Based Rare Plant Monitoring Using a Species Distribution Model and AI-Based Object Detection

Monitoring rare plant species is used to confirm presence, assess health, and verify population trends. Unmanned aerial systems (UAS) are ideal tools for monitoring plants because they can efficiently collect data without impacting the or endangering personnel. However, UAS flight planning be subjective, resulting in ineffective use of time overcollection imagery. This study a Maxent machine-learning predictive model create targeted areas monitor Geum radiatum, an endangered endemic Blue Ridge Mountains North Carolina. The was developed with ten environmental layers as predictors known locations training data. were derived from probability raster isolines delineated threshold based on parameters. Visual analysis imagery verified 33 discovered four previously undocumented occurrences. Semi-automated detection explored using neural network object detector. Although approach successful detecting on-ground images, no identified imagery, indicating that further improvements needed both acquisition computer vision techniques. Despite this limitation, presented research provides data-driven plan modeling, improving collection monitoring.