A Practical Method in Simultaneous Improvement of Spectral and Spatial Resolutions of Satellite Images in Urban Environments

One of the physical quantities that could be acquired from remote sensing data is surface reflectance in different regions of electromagnetic reflected spectra. Almost all the urban pixels that imaged by low/medium spatial resolution sensor systems represent composite radiance fields emanating from several distinct features with different reflectance''s within the sensor’s field of view. A general approach for describing urban environments is using classification methods and unmixing models، and among the useful models is spectral mixture analysis in which the mixed pixel reflectance is a linear combination of the reflectance and fractions corresponding to land cover types within the pixel. This model can be grouped to spectral unmixing and spatial unmixingnes. The objective of spatial unmixing models is to determine the reflectance spectrum of the classes in each mixed pixel. In this paper، spectral reflectance of important urban classes is determined using spatial unmixing model. To this end، spatial information from IKONOS imagery and spectral information from Hyperion data have been employed. Spatial resolutions of these images are 4m and 30m، and the numbers of bands are 4 and 242، respectively. The validity of the proposed method has been substantiated through comparison of original Hyperion image with the reconstructed image. Afterwards، using the obtained spectral reflectance، an image with 4m spatial resolution and 136 bands has been produced and variability of urban land covers has been taken into account.