A rapid assessment of Water Hyacinth (aquatic vegetation) mapping using machine learning for Circular Economic Opportunity in Sri Lanka

Water hyacinth, an invasive aquatic plant, has seriously infested the waterbodies of Sri Lanka and disturbed the freshwater aquatic environment. This study presents a rapid assessment of water hyacinth mapping in Sri Lanka using earth observation (EO) and machine learning techniques, highlighting its potential for circular economic opportunities. While posing significant environmental challenges, water hyacinth offers economic potential through its use in biofuel production, composting, bioremediation, and as biochar soil ameliorant. The methodology integrates Principal Component Analysis, K-means clustering and Otsu thresholding for precise water hyacinth area delineation. Initial results indicate a total biomass estimate of the weed to be $4111 \pm 584$ tons per year for the selected 28 water bodies. These water bodies are chosen for their location, size and reported water hyacinth invasion. The findings underscore the potential of integrating EO and machine learning into ecological management and offer a scalable model for similar assessments in other regions.