1 2004 
  
MPL. 
can be 
e 95% 
3 input 
and is 
Better 
29.08% 
r 99%. 
yrk has 
work. 
-- input 
— 2input | 
— 3input | 
—4input | 
--- Sinput | 
| 
| 
es 
broader 
veral oil 
tremely 
oil spill 
- 42:1 
^. There 
spill but 
oil spill 
oil spill 
types of 
ultilayer 
iges and 
uts to a 
i image 
F almost 
in local 
ation and 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004 
conjugate gradient seems to solve this problem. Moreover, 
MLP has smaller memory requirements for the classification 
and has better generalization than the RBF. 
  
  
  
  
  
  
  
b) Classified image 
  
  
  
Figure 8. Image result of the developed method 
From the two neural network models examined MLP works 
more reliably than RBF for oil spill detection. The mean 
performance for all RBF topologies examined was 77.62% 
while for MPL 98.98%. Several topologies were examined 
using the constructive method. The topology best suited for the 
classification procedure was the MLP 4:2:1 according to 
specific inputs. Classification accuracy was 99.433% for the 
above topology. The high performance of neural networks as 
classifiers was confirmed by producing accuracy 99,29 — 
99,60% when applied to other images, which contain oil spills 
and are captured under the same wind conditions. For RBF, the 
best performance achieved was 99.08% with 4:4:2 topology but 
the more reliable topology was topology with 3 inputs (3:3:2, 
3:4:2, 3:5:2) with a mean performance of 98.37%. 
Further examination is needed using images containing 
different sea states and different types of oils spills. Moreover, 
the performance of other neural network types like Support 
Vector Machines (SVM) and Recurrent networks (like 
Hopfield) need to be investigated. 
REFERENCES 
Alpers, W., Wismann, V., Theis, R., Huhnerfuss, H., Bartsch, 
N., Moreira, J., Lyden, J., 1991. The damping of ocean surface 
waves by monomolecular sea slicks measured by airborne 
multi-frequency radars during the saxon-fpn experiment. Proc. 
IGARSS'9] 
Anne, H., Schistad, S., Storvik, G., Solberg, R, Volden, E., 
1999, Automatic Detection of Oil Spills in ERS SAR Images, 
IEEE Transactions on geoscience and remote sensing, 37 (4), 
pp. 1916-1924 
729 
Benelli, G., Garzeili, A., 1999. Oil-Spills Detection in SAR 
Images by Fractal Dimension Estimation, /EEE IGARSS 1999 
Proceedings, pp. 218-220, Hamburg 
Bishop, C., 1995. Neural Networks for Pattern Recognition, 
Oxford University Press, Oxford 
Frate, Del. F., Petrocchi, A, Lichtenegger, J., Calabresi, G., 
2000. Neural Networks for oil spill detection using ERS-SAR 
data, /EEE Transactions on geoscience and remote sensing, 38 
(5), pp. 2282-2287 
Gade, M., Scholz, J., Viebahn, C., 2000. On the delectability of 
marine oil pollution in European marginal waters by means of 
ERS SAR imagery”, IEEE IGARSS 2000 Proceedings, VI, pp. 
2510-2512, Hawaii 
Kanellopoulos, L, Wilkinson, G., Roli, F., Austin, J., 1997. 
Neurocomputation in Remote Sensing Data Analysis, Springer 
Kavzoglu, T., Mather, P., 2003. The use of backpropagation 
artificial neural network in land cover classification, 
International Journal of Remote Sensing, 24 (23), pp 4907- 
4938 
Kubat, M., Holte R., Matwin, S., 1998. Machine learning for 
the detection of oil spills in satellite radar images, Machine 
Learning, 30 (2-3), pp. 195-215 
Lu, J., Lim, H., Liew, S., Bao M., Kwoh, L., 1999. Ocean Oil 
Pollution Mapping with ERS Synthetic Aperture Radar 
Imagery” IEEE IGARSS 1999 Proceedings, pp. 212-214 
Martinez, A. and Moreno, V., 1996. An Oil Spill Monitoring 
System Based on SAR Images, Spill Science & Technology 
Bulletin, 3 (1/2), pp. 65-71 
Pavlakis, P., 2001. On the monitoring of illicit vessel discharges 
using spaceborne sar remote sensing; a reconnaissance study in 
the Mediterranean Sea, Annals of Telecommunition, tome 56, 
nol 1/12, November-December 
Pavlakis, P., Sieber, A., Alexandrinou, S., 1996. On the 
Optimization of Spaceborne SAR Capacity in Oil Spill 
Detection and the Related Hydrodynamic Phenomena, Spill 
Science and Technology Bulletin, 3 (1/2), pp. 33-40 
SNNS, Stuttgart Neural Network Simulator,User Manual, 
Version 4.2, 1998, University Of Tubingen 
Topouzelis, K., Karathanassi, V., Pavlakis, P., Rokos, D., 2002. 
Oil Spill Detection: SAR Multi-scale Segmentation & Object 
Features Evaluation, 9th International Symposium on Remote 
Sensing - SPIE, 23-27 September Crete, Greece 
Topouzelis, K., Karathanassi, V., Pavlakis, P., Rokos, D., 2003. 
A Neural Network Approach to Oil Spill Detection Using SAR 
Data, 54" International Astronautical Congress, Bremen, 
Germany, 29.Sept - 03.Oct 
Ziemke, T., 1996. Radar image segmentation using recurrent 
artificial neural networks, Pattern Recognition Letters, 17, pp. 
319-334