Hyperspectral image analysis for oil spill detection: a comparative study Online publication date: Mon, 14-May-2018
by Sahar A. El-Rahman; Ali Hussein Saleh Zolait
International Journal of Computing Science and Mathematics (IJCSM), Vol. 9, No. 2, 2018
Abstract: In the last years, oil spill detection by hyperspectral imaging has been transferred from experimental to operational. In this paper, researchers attempted to use and compare four classification approaches for the identification of oil spills. The hyperspectral image classification approaches 'namely' are support vector machine (SVM), parallelepiped, minimum distance (MD) and binary encoding (BE). These approaches used to identify the oil spill areas in both two study areas which are selected as oil-spill areas in the Gulf of Mexico and the Adriatic Sea. The classifiers are applied to the study areas after pre-processing that include the spatial and spectral subset and atmospheric correction. Whereas, the classifiers applied to the full dataset and region of interest (ROI) before and after performing principal component analysis (PCA). The PCA is utilised to eliminate redundant data, reduce the vast amount of information and consequently, decrease the processing times. The findings indicate that the SVM, MD and BE approaches supply a high classification accuracy better than parallelepiped approach using both datasets obtained from both selected region.
Online publication date: Mon, 14-May-2018
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Computing Science and Mathematics (IJCSM):
Login with your Inderscience username and password:
Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.
If you still need assistance, please email firstname.lastname@example.org