Segmentation of EIT images using fuzzy clustering. A preliminary study

Vincent Barra (1) , Véronique Delouille (2) , Jean-François Hochedez (3) , Pierre Chainais (4)

1 LIMOS - Laboratoire d'Informatique, de Modélisation et d'Optimisation des Systèmes
2 Solar Physics Department [Brussels]
3 ROB - Royal Observatory of Belgium = Observatoire Royal de Belgique
4 LIMOS - Laboratoire d'Informatique, de Modélisation et d'optimisation des Systèmes

We present a new application of a signal processing technique called fuzzy clustering for automatic identification of various structures seen in EIT images. This technique gives for each pixel a probability of belonging to a particular class. By assigning each pixel to the class for which it has the greatest probability of belonging, we obtain image segmentations. In EIT 19.5 nm images we distinguish the Quiet Sun, Coronal Holes, and the Active Regions, whereas in EIT 30.4 nm we extract the plages and part of the network boundaries. We also show how a multiwavelength approach leads to an improved segmen-tation of the different coronal structures.

Mots clés: Irradiance study; EIT images; Fuzzy Cluster- ing; Multiwavelength analysis

Domaines: Traitement du signal et de l'image [eess.SP]

Vincent Barra, Véronique Delouille, Jean-François Hochedez, Pierre Chainais. Segmentation of EIT images using fuzzy clustering. A preliminary study. Proceedings of European SPM-11, Sep 2005, Leuwen, Belgium. ⟨hal-01622942⟩