Oil Spill Detection in Marine Environments Using AIS and Satellite Data

Overview

Detecting oil spills in marine environments using Automatic Identification System (AIS) data and satellite datasets integrated with advanced machine learning techniques.

Solution

A system for detecting oil spills in marine environments by combining AIS data and satellite datasets, leveraging machine learning.

• Machine Learning for Anomaly Detection: DBSCAN clustering algorithm on AIS data to identify anomalies such as unexpected vessel behavior.
• Customized CNN for Oil Spill Detection: A Convolutional Neural Network trained on Synthetic Aperture Radar (SAR) images to detect oil spills.
• Integrated System: AIS and satellite data for precise detection.

Technical Approach

Technologies Used

• Programming Languages: Python
• Frameworks and Libraries:
  • Scikit-learn for DBSCAN
  • TensorFlow/Keras for CNN
  • Pandas and NumPy for data handling
  • OpenCV for image manipulation
  • SNAP tool for SAR image processing
• Database: MongoDB

Process

1. Data Ingestion: AIS data and satellite imagery.
2. Preprocessing: Handling missing values and preparing datasets.
3. Model Development:
   • DBSCAN for anomaly detection in AIS data.
   • CNN for analyzing SAR images.
4. Integration: AIS anomaly detection results validated by satellite image analysis.