Urban Land Cover Mapping Using Object/pixel-based Data Fusion and Ikonos Images

The use of high resolution IKONOS imagery should make it possible to extract urban features like buildings and roads more easily than conventional satellite image data. This work examines the synergetic ability of using spectral information from multispectral data in combination with spatial information from panchromatic data in urban land cover mapping using IKONOS Imagery. In this work we tried to answer the question, how much data fusion influences the classification results at the pixel and object level? The work contains some main steps. First, the land cover classification based on the Object Oriented Classification (OOC) and Maximum Likelihood Classification (MLC) algorithms using multispectral IKONOS image has been carried out. Second, the panchromatic image was fused with multispectral data using wavelet data fusion and overly algorithms. Finally, land cover classification based on the wavelet fused images using MLC and OOC algorithms has been undertaken. The accuracy of the results was calculated and they show that the increase of spatial resolution in fused data has a positive influence on the results.