Monitoring Water Changes using Sentinel-2 Imagery and Machine Learning

The pressures placed on freshwater resources by human activity and climate change are growing. For efficient resource management and conservation initiatives, monitoring changes in water bodies is essential. This work tackles the problem of precisely identifying and categorizing water body dynamics by utilizing Sentinel-2 satellite imagery and Google Earth Engine (GEE). Water bodies are often difficult to distinguish using traditional land cover classification methods, especially when dealing with gradual changes over time. Resolution suggests utilizing a Support Vector Machine (SVM) classifier trained on Sentinel-2 data from two different time periods as part of a machine learning-based strategy. Furthermore, changes in the presence of water are detected using the Normalized Difference Water Index (NDWI), which offers a reliable method of change detection. This study shows that the suggested methodology is effective in classifying land cover, including water bodies, with a high classification accuracy of 97%. Detailed maps that show variations in water extent over time can be produced by combining Sentinel-2 data, GEE, and machine learning approaches. Advantages The results enhance environmental monitoring capacities and support well-informed decision-making for the protection and sustainable management of water resources. This research responds to need for monitoring off water bodies by combining Sentinel-2 imagery, sophisticated pre-processing, the NDWI index, and Support Vector Machine classifier on Google Earth Engine. Through the automated detection process and visualization of results using color mapping, this research offers an efficient method for identifying the changes in water level.