Detecting Oil Spills at Marine Environment Using Automatic Identification System (AIS) And Satellite Datasets

Aim:

          To design and implement a high-accuracy, real-time oil spill detection system that integrates Automatic Identification System data and satellite imagery using the advanced YOLOv11 deep learning model for environmental monitoring and rapid disaster response.

Abstract:

       This project proposes an intelligent oil spill detection framework using YOLOv11, an advanced deep learning model, integrated with Automatic Identification System data and satellite imagery. The proposed system improves real-time identification and segmentation of oil spills in marine environments. The system correlates AIS ship data with satellite imagery to identify potential spill sources and track vessel activity around contaminated areas. It provides real-time visualization, reporting, and through a web interface. Experimental results show that YOLOv11 achieves up to high detection accuracy, surpassing earlier YOLO -based model demonstrating its robustness in complex marine environments.

Proposed System:

          The proposed system employs an advanced YOLOv11 deep learning model for real-time oil spill detection and segmentation from satellite imagery. The YOLOv11 model is trained on a customized, high-quality oil spill dataset consisting of various oceanic conditions to ensure robustness and generalization. It automatically identifies and segments oil spill regions with high precision, enabling quick and reliable detection even under complex environmental factors such as waves, reflections, and lighting variations. By leveraging YOLOv11’s improved architecture and enhanced feature extraction capabilities, the system achieves outperforming earlier versions like YOLOv11. The fully automated framework provides real-time visualization, alert generation, and data reporting through efficient monitoring and faster response in marine environmental management.

Advantage:

1. Improved detection accuracy.
2. Enhanced real-time performance suitable for practical applications.
3. Lightweight and efficient model design for ease of deployment.