Oil Spill Detection and Verification in Nothern Bintan

Putri Suhendi, Brigitta Aurelia (Department of Statistical Computing, Politeknik Statistika STIS, Jakarta, Indonesia)
Marsisno, Waris (Department of Statistical Computing, Politeknik Statistika STIS, Jakarta, Indonesia)

Zero : Jurnal Sains, Matematika, dan Terapan
Vol 9, No 2 (2025): Zero: Jurnal Sains Matematika dan Terapan

The Riau Islands Province, particularly northern Bintan Island, is strategically located near countries such as Singapore, making it a gateway to regional and international markets but also vulnerable to oil spills. This study aims to detect oil spill areas using GLCM texture analysis, adaptive thresholding, and various machine and deep learning models, followed by look-alike verification. The XGBoost model achieved the best performance with an accuracy of 0.9772, detecting oil spill areas of 5,400,241 m² and look-alike regions covering 1,333,045 m² on March 23, 2024. The findings also indicate that inland waters are often misidentified as spills, highlighting the importance of verification. This study is the first to integrate several of these methods for Sentinel-1 based oil spill detection in Bintan waters, as a new approach to accurate and efficient regional monitoring.