A NEW MERGING METHOD FROM MULTISPECTRAL AND PANCHROMATIC SPOT IMAGES FOR A MAPPING OF FLOODPLAIN LAND-COVER

The different methods to merge multiresolution and multispectral data are becoming currently used to study the environmental components. Actually, the high spectral resolution of the SPOT XS data combined to the hight spatial resolution of the SPOT panchromatic data allow a better analysis of the land cover (in this case an Alpine floodplain is studied). The currently merging methods (P+XS, HIS...) are not satisfying; they are distorting the spectral characteristics of the XS images and, in result the photointerpretation becomes difficult. A method allowing to simulate 10-m resolution XS-data, while conserving the spectral properties of original 20-m data is presented. This method is using a multiresolution analysis procedure based upon the wavelet transform; it is applied to a remotely sensed image SPOT P and SPOT XS of the river junction ''Arc-Isere'' (France).