Coastline Detection from Sentinel -۲ Imagery Using Machine Learning with Minimal Basepoint Input

This study presents a method for extracting coastline boundaries from Sentinel -۲ satellite images using the Support Vector Machine (SVM) classification technique, with a focus on the Khezershahr coastline. Sentinel -۲ provides high-resolution, multispectral data that is essential for detailed coastal monitoring. The SVM, a powerful supervised learning algorithm, was selected for its capability to handle complex and high-dimensional datasets. It was trained on visual spectral bands to distinguish between land and water during the preprocessing stage. The resulting coastline outlines were validated against high resolution reference data, confirming the method’s accuracy and reliability. This technique shows great promise for ongoing, automated monitoring of coastal changes, aiding in more effective coastal management and decision-making.

کلیدواژه ها:

Shoreline ، SVM ، Machine Learning