The examples provided in this paper enhance the capability and the benefits of the hierarchical description of the information provided but MRA and WT for data fusion, In each case, the results achieved where evaluated and compared to methods already in use. These comparisons demonstrate the benefits of the multiresolution models derived from the MRA representation. The models derived from this representation are of various complexity and are based on the statistics of the wavelet coefficients images. Some other classes of models have been defined. Basseville et al. [12, 13] have linked the multiscale representation of information in an image with stochastic processes through the definition of multiscale autoregressive models. As an example in the field of RS, Fosgate et al. [14] have demonstrated the efficiency of these models for the segmentation of SAR images and the detection of anomalies in these images. New classes of models are under definition and some links between the geostatistics field and the wavelet are emerging. More and more applications in the field of RS take advantage of this hierarchical description of the information provided by MRA and WT. The development of the modeling based on WT in the field of RS will lead to a better understanding of the natural processes observed from satellites.
展开▼
