Comparative Analysis between ResNet Models on Marine Oil spill detection

Joseph, Abin (2023) Comparative Analysis between ResNet Models on Marine Oil spill detection. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (776kB) | Preview
Preview	PDF (Configuration Manual) Download (316kB) | Preview

Abstract

Marine oil spills remain a significant threat to ecosystems and maritime safety, necessitating advanced detection methodologies. Despite a decrease in the frequency of spills, challenges exist in timely and accurate detection. This research evaluates state of the art deep learning models, including ResNet50, ResNet50V2, and ResNet101, with the previously used VGG19 model for SAR image classification in oil spill detection. The study addresses misclassification issues, focusing on the effectiveness of these models in classifying SAR images with oil-like and non-oil-like features. The evaluation utilizes metrics such as the classification report, confusion matrix, and ROC curve. Results shows that ResNet50 outperforms other models, achieving a weighted F-score of 0.95 and a ROC curve area of 0.99. The research contributes valuable insights to environmental monitoring, emphasizing the potential replacement of VGG19 with ResNet50 for improved oil spill detection.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Shahid, Abdul UNSPECIFIED
Uncontrolled Keywords:	Marine oil spills; SAR imagery; Deep learning models; ResNet50; Classification evaluation; Environmental monitoring
Subjects:	G Geography. Anthropology. Recreation > GE Environmental Sciences Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning H Social Sciences > HD Industries. Land use. Labor > Specific Industries > Oil Industry
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	09 May 2025 10:01
Last Modified:	09 May 2025 10:01
URI:	https://norma.ncirl.ie/id/eprint/7536

Actions (login required)

View Item