Mapping ash trees from WorldView-2 and 3 data

This project was motivated by the threat caused by the emerald ash borer, which is spreading across several regions in North America. Mapping of ash trees is an important step in the preparation for a possible infestation of this frightening insect pest, which can destroy endemic ash stands. In this research, the potential of WorldView-2 and 3 images was analyzed to identify ash trees. The study area is located in Sherbrooke (Eastern Quebec, Canada). The region is not infested by the emerald ash borer yet. In addition to remote sensing images, a large GIS database, with the localization of individual trees and their identification, was provided by the City of Sherbrooke to support the project. The final results were presented as probability maps of ash detection. Using a threshold >=0.7 (70% confidence), a global identification ratio of 83% was found with Worldview-2 image, compared to the City ground-based inventory.