Work

Detección de manchas de petróleo en imágenes SAR mediante mintree y Deep Learning

Year: 2020
Type: dissertation
Abstract: Oil spills constitute one of the main risk factors for oceanic, maritime and coastal ecosystems. They cause a problem that can be catastrophic for local marine fauna and the growth of maritime commerce. Complementing the thesis of the students Ismael Vivó and Olga Ruiz, this project proposes an investigation based on oil spill detection problem in SAR images (Synthetic-Aperture Radar) by means of Image Processing Methods, together with Machine Learning classification techniques in the context of Semantic Segmentation applications. This work applies the learning algorithms SVMs (Support Vector Machines) and CNN (Convolutional Neural Network) on the min-tree representation of the images, or on the images directly. Finally, the results are evaluated through the metrics used in the context of detection such as Precision, Recall and F-Score, and compared with the obtained in the previous theses.
Language: Spanish
Subfield: Pollution
Open Access status: green
OpenAlex ID: W3011970031
Cites: 6
Cited by:
Related to: 9