Water Hyacinth (Eichhornia crassipes) Detection Using Coarse and High Resolution Multispectral Data

Efficient detection and monitoring procedures of invasive plant species are required. It is crucial importance to deal with such plants in aquatic ecosystems, since they can affect biodiversity and, ultimately, ecosystem function services. In this study, it intended detect water hyacinth (Eichhornia crassipes) using multispectral data different spatial resolutions. For purpose, high-resolution (<0.1 m) acquired from an unmanned aerial vehicle (UAV) coarse-resolution (10 Sentinel-2 MSI were used. Three areas a high incidence located the Lower Mondego region (Portugal) surveyed. Different classifiers used perform pixel-based both datasets. From used, results achieved by random forest stand-out (overall accuracy (OA): 0.94). On other hand, support vector machine performed worst (OA: 0.87), followed Gaussian naive Bayes 0.88), k-nearest neighbours 0.90), artificial neural networks 0.91). The higher resolution UAV-based enabled us small amounts hyacinth, which could not be detected data. However, despite coarser resolution, satellite analysis identify coverage, compared well survey. Combining datasets even considering resolutions, was possible observe temporal evolution hyacinth. This approach proved effective way assess effects mitigation/control measures taken study areas. Thus, applied environments monitor their changes over time.