Comparison of four different methods to merge multisensor and multiresolution satellite data for the purpose of mapping

Kurzfassung

Most satellite data have either a high spectral resolution or a high spatial resolution. Numerous methods have been developed to merge thecolour information of a higher spectral resolution data set with the spatial information of a higher spatial resolution data set. However,the merging effects of the methods vary in different studies and areas. In this paper, the four most recommended and most effective methods of IHS (Intensity, Hue, Saturation), PCS (Principal ComponentSubstitution), RVS (Regression Variable Substitution) and SVR (Synthetic Variable Ratio) were assessed. The TM data and SPOT pan data of four different large areas were used in the merging process. Visual and graphic comparisons demonstrate that the data merged with the SVR method reproduce the colour information of the original TM data best. The spatial resolution of the data merged with the IHS, PCS and SVR methods are similar. But the spatial details can be visually better observed in the IHS and PCS data than in the SVR data. The spectral and spatial effects of the RVS method are the worst. For the purpose of land use classification the SVR method os more suitable. However, for the purpose of image mapping the IHS and PCS methods show a better merging effect.