Tree Crown Detection using UAV-captured High-Resolution Aerial Images for Baghdad University Campus

Abstract

The research paper's objective is to develop an automated system that uses high-resolution aerial images captured by Unmanned Aerial Vehicles (UAVs) to accurately detect and analyze tree crowns on the Baghdad University campus. This analysis of tree crowns is crucial for assessing vegetation health and distribution in urban areas, with applications in urban planning, ecology, and environmental studies. The proposed system utilizes computer vision and machine learning techniques to detect and delineate tree crowns, facilitating objective analysis of tree canopy coverage. The study employed a Phantom Four drone flying at a height of 55 meters with an 80% overlap in image coverage. Ground control points were established using the Topcon_HIPER2 device and Static GPS Surveying Techniques. The flight plan was designed with PIX4Dcapture software, and the resulting ortho-map had a 2 cm resolution. The research used the YOLOv3 artificial intelligence technique for tree detection after training a customized dataset. The model achieved impressive performance metrics, including a low loss function value of 0.023, a high SDC value of 0.984, and a robust Jaccard Index value of 0.943.