DETECTING AND MAPPING OF WATER HYACINTH IN LAKE VICTORIA USING RADAR POLARIMETRIC DATA

Water hyacinth (WH) is one of the dreadful noxious invasive species whose origin is traced to South America. Compared to in-situ measurements, remote sensing offers a less expensive way of monitoring WH. In this research, the Optimization of Power Difference (OPDiff) change detection technique was used in mapping WH. Sentinel-1 IW SLC data from 2017 to 2022 were analyzed to get polarimetric covariance matrices using the VV, VH and VV*VH. Preprocessing, co-registration and detection steps were carried out using the GPT processor of ESA SNAP graph builder. The OPDiff applied in the study areas enabled the detection and extraction of spatial extent of WH in single image and in time series by identifying the changes occurring on the minimum eigenvalues. When evaluated against Sentinel-2 data, the algorithm recorded on average 92.6% & 89.9% precision and accuracy respectively. Heatmaps were generated for the study area and demonstrated variability of WH in time and space. This research demonstrates the capability of using polarimetric radar data to detect and monitor WH.