MKLab presents the Oil Spill Detection Dataset to the research community
(a) Sample of a SAR image and (b) Corresponding annotated image. Cyan color corresponds to oil spills, red to look-alikes, brown to ships, green to land and black is for sea surface.
Aiming to enhance the research activity in the field of oil spill detection, the Multimedia Knowledge and Social Media Analytics Laboratory (MKLab) has developed the Oil Spill Detection Dataset, a dataset containing Synthetic Aperture Radar (SAR) images depicting oil spills and other relevant instances. Current dataset is part of our research work presented in the paper entitled as “Oil spill identification from satellite images using deep neural networks” and could be beneficial for remote sensing applications mostly. The dataset is available to the research community through a specific access request process. Currently, more than 50 institutions and organizations have utilized the Oil Spill Detection Dataset in research projects, which among of them are:
- Greenpeace UK – Development of an automatic oil spill detection tool based on SAR data with a web interface.
- National Space Science Agency - Automatic monitoring system for oil spill detection in the Arabian gulf area
- University of Minnesota - Detecting oil spills on water surfaces using deep learning.
- Polytechnic University of Catalonia - Oil Spill detection on SAR image with mintree representation and machine learning
- National Technical University of Athens - Automatic detection of oil pollution from remote sensing data in the Mediterranean Sea