Chapter Mapping Yucca gloriosa in coastal dunes: evaluating the cost and time efficiency of photointerpretation, machine learning and field detection approaches

Abstract

Biological invasions threaten biodiversity and cause significant economic and ecological costs. Effective management of invasive species is crucial, as highlighted by the European Community's Regulation 1143/2014 on Invasive Alien Species (IAS). This study focuses on coastal dune ecosystems, particularly assessing the time and cost-effectiveness of three monitoring methods for detecting and mapping alien plants: photointerpretation, machine learning classification, and field monitoring. Yucca gloriosa L., an invasive species in Regional Park of Migliarino-San Rossore-Massaciuccoli (Tuscany, Italy), served as the target species. Using RGB DJI Phantom 4 Pro v. 2.0 and DJI P4 Multispectral drones, images were analyzed via photointerpretation and machine learning. Photointerpretation, though precise, was time-consuming and subjective. Machine learning minimized human effort but required extensive computing. Field monitoring produced accurate maps but was labor-intensive and limited by accessibility issues. This study concludes that UAV-based monitoring of Y. gloriosa is optimal for balancing cost and time efficiency in coastal dune ecosystems.