Wavelet and evidence theory for object-oriented classification: Application to change detection in Rennes metropolitan area

International audienceThis paper is concerned with the estimation of the dominant orientation of textured patches that appearin a number of images (remote sensing, biology or natural sciences for instance). It is based on the maximizationof a criterion that deals with the coefficients enclosed in the different bands of a waveletdecomposition of the original image. More precisely, we search for the orientation that best concentratesthe energy of the coefficients in a single direction. To compare the wavelet coefficients between the differentbands, we use the Kullback–Leibler divergence on their distribution, this latter being assumed tobehave like a Generalized Gaussian Density. The space–time localization of the wavelet transform allowsto deal with any polygon that may be contained in a single image. This is of key importance when oneworks with (non-rectangular) segmented objects. We have applied the same methodology but usingother criteria to compare the distributions, in order to highlight the benefit of the Kullback–Leibler divergence.In addition, the methodology is validated on synthetic and real situations and compared with astate-of-the-art approach devoted to orientation estimation