Adequacy of the Gamma statistical distribution for low and high resolution sar images, using the kullback relative entropy - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2007

Adequacy of the Gamma statistical distribution for low and high resolution sar images, using the kullback relative entropy

Résumé

Synthetic Aperture Radar (SAR) imaging systems become more and more precise and some very high resolution (less than a meter) data are now available. Although very interesting, notably for earth monitoring, SAR images are very noisy. This noise is usually modeled with fully developed speckle noise, leading to a Gamma distribution for the intensity of the pixels. Although this very simple model has demonstrated good behavior for low resolution SAR images (about 10 meter resolution), one can wonder if this model is still interesting for high resolution SAR images. In this paper, we use the Kullback relative entropy to define a degree of adequacy to Gamma distribution for the pixel intensities. Then we will be able to show that, provided one considers homogeneous regions, this degree of adequacy can be useful to see whether or not the pixel grey levels can be modeled with Gamma laws in low and high resolution SAR images.
Fichier non déposé

Dates et versions

hal-00165730 , version 1 (27-07-2007)

Identifiants

  • HAL Id : hal-00165730 , version 1

Citer

Frédéric Galland, Philippe Réfrégier, Gemma Pons Bernad. Adequacy of the Gamma statistical distribution for low and high resolution sar images, using the kullback relative entropy. PSIP 2007, Feb 2007, Mulhouse, France. ⟨hal-00165730⟩
63 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More