Fusion of Airborne SAR Image and Color aerial image for Land Use Classification

摘要

The fusion of multi-source data is one of the most promising techniques for improved classification of remote sensing images. This paper utilized the images by fusing the high-resolution optical and SAR images for land use classification. The high-resolution optical and SAR images present the land features for many aspects and supplement information to each other because of their different imaging modes and wavebands , so classical methods such as Hue - Intensity -Saturation(HIS) transform based, Brovey(color normalized),PrincipaI Component Substitution(PCS) approaches and wavelet-based method are used in fusing process, then the fused results been evaluated through the calculation of some quantitative index(Mean grey value, Standard error, Entropy, Definition). The wavelet-based fused image shows better effect than others. A MLC(maximum likelihood classification) method was employed to the wavelet-based fused image .Classification accuracy was assessed using high-resolution aerial orthophotos. The overall accuracy for six classes(Settlement, River, Agricultural field, Country road, Tree nurseries, Bare land) was found to be 94.36% with kappa coefficient of 0.92.