Gold Medal winner in the Bachelor’s and Master’s Degree category:
#clge #deeplearning #innovation #automation
1. First Prize – Mohamed Al Hashmi – Automating the detection of oil spills from Sentinel-1 SAR imagery using deep learning (Geospatial Surveying and Mapping at Newcastle University, United Kingdom).

Oil spills are described as the discharge of liquid petroleum hydrocarbons
into the marine environment because of human activities (Li, 2016). The
rising global demand for crude oil has led to the expansion of marine oil
transportation, increasing the risk of oil spills. It causes severe harm to the
marine ecosystem and nearby coastal areas, leading to significant
environmental and economic consequences (De Kerf et al., 2020). Oil spills
contaminate water sources, affect the availability of clean water, cause
desalination plants to shut, and damage mangroves, beaches and aquatic
animals (Anselain et al., 2022). It could happen in freshwater bodies or
oceans. Authorities desperately need reliable and automated oil spill
monitoring systems to enable effective response and mitigation efforts, as
well as monitoring of regulatory compliance by ships (De Kerf et al., 2020).
Moreover, minimising the damage caused by oil spills will contribute to
achieving the United Nations Sustainable Development Goal 6, introduced
to improve water quality, and Goal 14, designed to protect marine life.
Manually extracting oil spills from Synthetic Aperture Radar (SAR) imagery
can be time-consuming because the process involves visual inspection of
large amounts of data. Machine learning is vital for oil spill detection
because it can improve the speed and efficiency of detecting oil spills. Deep
learning (DL) offers advantages over traditional machine learning methods
in that it can automatically learn the features to extract to solve a task from
the training data itself. It can also handle large amounts of data more
efficiently and accurately than manual methods. This project aimed to
develop an automated system for oil spill detection using deep learning and
Sentinel-1 SAR imagery to enable faster, more accurate and reliable
detection and response to oil spills.

[ Ссылка ]

свернуть