Komeil Rokni - Department of Geomatic Engineering, Gonbad Kavous University, Golestan, Iran,

In this study, the effectiveness of different transforming techniques of wavelet resolution merge is evaluated for fusion of ASTER-L1B and RADARSAT-1 SAR data. In doing so, the SB (Single Band), IHS (Intensity, Hue, Saturation) and PC (Principal Component) techniques are applied to merge the input images. Since suitability of the fusion algorithms for various applications depend on the quality of the fused images, several indicators are used to assess performance of the applied techniques. The results of standard deviation difference, relative variation difference and root mean square error analyses indicate advantage of the SB transforming technique comparing to the PC and HIS techniques, while the results of relative mean difference show superiority of the PC technique. The results also indicate that IHS is the worst technique based on analyses of the standard deviation difference, relative variation difference and root mean square error, while the output of SB is the worst image based on the relative mean difference analysis. Finally, based on the achieved results, it can be concluded that all three transforming techniques are well preserved spectral quality of the original multispectral image, but the output of SB technique indicate slightly advantage comparing to the outputs of PC and HIS techniques for the fusion of ASTER and RADARSAT SAR data. In conclusion, the results revealed effectiveness of wavelet resolution merge approach for multi-sensor data fusion.

ASTER, RADARSAT SAR, image fusion, wavelet resolution merge