Detection and Monitoring of Alien Weeds Using Unmanned Aerial Vehicle in Agricultural Systems in Sardinia (Italy)

Abstract

Emerging technologies such as high-resolution Unmanned Aerial Vehicles (UAVs) surveys combined with object-based image analysis, and field surveys could represent a reliable, precise, and effective tool to support land management in agricultural systems. The technological advances of UAVs can also promote the detection and regular monitoring of invasive alien plants and agricultural weeds. The objective of the study has been to identify, map and monitor alien weed species in agricultural systems to provide an overview of the future applications and challenges of precision farming. In particular, we evaluated how UAV imagery can be used to assess the cover of Oxalis pes-caprae, present in several crops in Sardinia as an alien invasive weed, with negative direct and indirect effects on the affected crops. Our core assumption is that the most reliable species discrimination can be achieved by targeting flights during flowering to allow easier detection due to species-specific spectral differences. Therefore, O. pes-caprae infestation was acquired using RGB camera installed on board a Phantom 4 pro. As a result, we presented the mapping of O. pes-caprae, highlighting the cost-effectiveness and replicability of this approach to detect the presence of this alien weed in agricultural fields.