XG-DeepLabV3+: Revolutionizing Ocean Eddy Detection with Satellite Imagery

Eddies are circular water movements that swirl against the main flow, causing turbulence and mixing. These dynamic ocean features play crucial roles in global energy transfer, nutrient distribution, and marine habitats. However, identifying oceanic eddy remains challenging due to local factors like turbulence and boundaries. Traditional algorithms are less effective compared to datasets like Argo floats, which measure ocean temperature and salinity. This work proposes the XG-D <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$e$</tex> epLabV3+ algorithm, which combines XGBoost and DeepLabV3+models to enhance the accuracy of detecting oceanic eddies by using satellite remote sensing SST data DeepLabV3+extracts feature from TIFF-format raster data into NumPy arrays, then input these features into XGBoost to optimize the residuals and update the model. Experiment results demonstrate that XB-DeepLab V3+ effectively improves precision for detecting small-scale vortices but only identifying 28% of the eddies detected by XGBoost under the same region.