Share Email Print
cover

Proceedings Paper

Measuring marine oil spill extent by Markov Random Fields
Author(s): Miguel Moctezuma; Flavio Parmiggiani; Ludwin Lopez Lopez
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

The Deepwater Horizon oil spill of the Gulf of Mexico in the spring of 2010 was the largest accidental marine oil spill in the history of the petroleum industry. An immediate request, after the accident, was to detect the oil slick and to measure its extent: SAR images were the obvious tool to be employed for the task. This paper presents a processing scheme based on Markov Random Fields (MRF) theory. MRF theory describes the global information by probability terms involving local neighborhood representations of the SAR backscatter data. The random degradation introduced by speckle noise is dealt with a pre-processing stage which applies a nonlinear diffusion filter. Spatial context attributes are structured by the Bayes equation derived from a Maximum-A-Posteriori (MAP) estimation. The probability terms define an objective function of a MRF model whose goal is to detect contours and fine structures. The markovian segmentation problem is solved with a numerical optimization method. The scheme was applied to an Envisat/ASAR image over the Gulf of Mexico of May 9, 2010, when the oil spill was already fully developed. The final result was obtained with 51 recursion cycles, where, at each step, the segmentation consists of a 3-class label field (open sea and two oil slick thicknesses). Both the MRF model and the parameters of the stochastic optimization procedure will be provided, together with the area measurement of the two kinds of oil slick.

Paper Details

Date Published: 14 October 2014
PDF: 6 pages
Proc. SPIE 9240, Remote Sensing of the Ocean, Sea Ice, Coastal Waters, and Large Water Regions 2014, 92400A (14 October 2014); doi: 10.1117/12.2068325
Show Author Affiliations
Miguel Moctezuma, Univ. Nacional Autónoma de México (Mexico)
Flavio Parmiggiani, Istituto di Scienze dell'Atmosfera e del Clima, CNR (Italy)
Ludwin Lopez Lopez, Univ. Nacional Autónoma de México (Mexico)


Published in SPIE Proceedings Vol. 9240:
Remote Sensing of the Ocean, Sea Ice, Coastal Waters, and Large Water Regions 2014
Charles R. Bostater; Stelios P. Mertikas; Xavier Neyt, Editor(s)

© SPIE. Terms of Use
Back to Top