A comparative study of fully automatic and semi-automatic methods for oil spill detection using Sentinel-1 data.

Area of Science:

• Environmental Science
• Remote Sensing
• Machine Learning

Background:

• Oil spills pose significant threats to marine ecosystems and local communities.
• Effective monitoring and rapid response are crucial for mitigating ecological damage.

Purpose of the Study:

• To assess oil spill detection and impact in Banten Province using Sentinel-1 satellite data and machine learning.
• To evaluate the performance of different machine learning classifiers for oil spill identification.

Main Methods:

• Utilized Sentinel-1 Synthetic Aperture Radar (SAR) data with VV polarization and an oil spill threshold of -25 dB.
• Applied advanced image processing, binary masking, and vectorization for GIS integration.
• Employed Artificial Neural Networks (ANN), Random Forest (RF), and Extreme Gradient Boosting (XGBoost) for oil spill classification.

Main Results:

• Temporal analysis revealed significant variability in spill sizes, with peaks on May 16 (79.686 km²) and July 3 (41.593 km²).
• Wind pattern analysis provided insights into spill dispersion dynamics.
• ANN showed superior discriminative ability (AUC=0.92), while RF achieved high accuracy (99.01%) and precision (99.02%).

Conclusions:

• The integrated approach of remote sensing, image processing, and supervised learning is viable for environmental monitoring.
• This study provides crucial data for minimizing ecological impacts and optimizing maritime disaster response plans.
• Advanced technologies are essential for combating ecological threats and protecting marine ecosystems.