Analysis of changes in quality assessment with scale

Claire Thomas (1) , Lucien Wald (2)

1 CEP - Centre Énergétique et Procédés
2 CEP/Sophia

Our work deals with the assessment of the quality of multimodal images synthesized at a better spatial resolution by the means of another image having such a resolution. In absence of reference images, the current protocols recommend to degrade spatially both sets of images and to perform fusion on these two new sets, thus producing synthesized images at the original low resolution. The quality budget is drawn at this scale and it is assumed that it is close to or worse than the one, which would have been drawn at high resolution if the reference images were available. This extrapolation hypothesis is the major limitation of the application of protocols. The validity of this hypothesis is studied for 44 case studies. Several distances are selected. They are compounded into several quality budgets, each of them giving an overall idea on the quality. We analyze changes in quality budgets with scales. It is found that the extrapolation hypothesis is validated in most cases. Nevertheless, the need for further investigation is underlined.

Mots clés: pan-sharpened; quality budget; ERGAS; multiresolution analysis; extrapolation; image fusion; satellite images; synthesis of multi-modality images; pan-sharpened.

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

Claire Thomas, Lucien Wald. Analysis of changes in quality assessment with scale. Fusion 2006, Jul 2006, Florence, Italy. pp.147. ⟨hal-00395066⟩