Detection of Moroccan Coastal Upwelling in SST images using the Expectation-Maximization

Ayoub Tamim (1) , Khalid Minaoui (2) , Khalid Daoudi (3) , Abderrahman Atillah (4) , Driss Aboutajdine (2)

1 LRIT
2 LRIT - Laboratoire de Recherche Informatique et Télécommunications
3 GeoStat - Geometry and Statistics in acquisition data
4 CRTS - Centre Royal de Télédétection Spatiale

This paper proposes an unsupervised algorithm for automatic detection and segmentation of upwelling region in Moroccan Atlantic coast using the Sea Surface Temperature (SST) satellite images. This has been done by exploring the Expectation-Maximization algorithm. The good number of clus- ters that best reproduces the shape of upwelling areas is selected by using the two popular Davies-Bouldin and Dunn indices. Area opening technique is developed that is used to remove and discarded the residuals noise in offshore waters not belonging to the upwelling region. The complete system has been validated by an oceanographer using a database of 30 SST images of the year 2007, demonstrating its capability and robustness for precise detection of Moroccan coastal upwelling.

Mots clés: Upwelling; Sea Surface Temperature; Expectation-Maximisation; Davies-Bouldin index; Dunn index; Area opening

Domaines: Traitement du signal et de l'image [eess.SP]; Interfaces continentales, environnement; Océan, Atmosphère

Ayoub Tamim, Khalid Minaoui, Khalid Daoudi, Abderrahman Atillah, Driss Aboutajdine. Detection of Moroccan Coastal Upwelling in SST images using the Expectation-Maximization. 10th International Symposium on Visual Computing, Dec 2014, Las Vegas, United States. ⟨hal-01120901⟩