Identifying and classifying water hyacinth (Eichhornia crassipes) using the HyMap sensor

In recent years, the impact of aquatic invasive specieson biodiversity has become a major global concern. In the Sacramento- San JoaquinDelta region in the Central Valley of California, USA, dense infestations of the invasive aquaticemergent weed, water hyacinth( Eichhornia crassipes) interfere with ecosystem functioning. This silent invader constantly encroaches into waterways, eventually making them unusable by people and uninhabitable to aquaticfauna. Quantifying and mapping invasive plant species in aquatic ecosystemsis important for efficient management and implementation of mitigation measures. This paper evaluates the ability of hyperspectral imagery, acquired using the HyMapsensor, for mapping water hyacinthin the Sacramento- San JoaquinDelta region. Classification was performed on sixty-four flightlines acquired over the study site using a decision tree which incorporated Spectral Angle Mapper (SAM) algorithm, absorption feature parameters in the spectral region between 0.4 and 2.5µm, and spectral endmembers. The total image dataset was 130GB. Spectral signaturesof other emergent aquaticspecies like pennywort ( Hydrocotyle ranunculoides) and water primrose ( Ludwigia peploides) showed close similarity with the water hyacinthspectrum, however, the decision tree successfully discriminated water hyacinthfrom other emergent aquaticvegetation species. The classification algorithm showed high accuracy (κ value = 0.8) in discriminating water hyacinth.