You must have a javascript-enabled browser and javacript and stylesheets must be enabled to use some of the functions on this site.


Bayesian Adaptive Oil Spill Segmentation of SAR Images via Graph Cuts

Sónia Pelizzari(1) and José Bioucas Dias(1)

(1) Instituto de Telecomunicações, Instituto Superior Técnico,Av. Rovisco Pais, 1049-001 Lisbon, Portugal


This paper presents a new Bayesian semisupervised segmentation algorithm aimed at oil spill detection in SAR images, a crucial step in any SAR based automatic oil spill surveillance system [1]. The data term, i.e., the density of the observed pixel amplitude given the region is modeled as finite mixture of gamma distributions [2]. This renders robustness to scatter fluctuations inside each region. The prior is an M-level Markov Random Field defined on a 2D grid enforcing local continuity in a statistical sense [3]. We adopt the maximum a posteriori (MAP) segmentation criterion. The optimization problem we are led to is solved efficiently by means of recent graph-cut techniques [4]. The effectiveness of the proposed method is illustrated with real ERS and ASAR data.


1. A. Solberg, G. Storvik, R. Solberg and E. Volden. Automatic Detection of Oil Spills in ERS SAR Images, IEEE Transactions on Geoscience and Remote Sensing, Vol. 37, No 4, July 1999. 2. M. Figueiredo, J. Leitão and A. Jain, On Fitting Mixture Models, In Energy Minimization Methods in Computer Vision and Pattern Recognition, pp 54-69, Springer-Verlag 1999. 3. J. Besag, On the statistical analysis of dirty pictures, Journal of the Royal Statistical Society B, 48(3):259-302, 1986. 4. V. Kolmogorov and R. Zabih. What Energy Functions Can Be Minimized via Graph Cuts? In IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol 26, No.2, February 2004.


Full paper


  Higher level                 Last modified: 07.10.03