Change detection in urban area by independent component analysis

The basic concept in using Remote Sensing data for change detection is that changes in land cover result in changes in radiance values that are more with respect to radiance changes caused by other factors such as illumination conditions and atmospheric conditions. Independent Component Analysis (ICA) is a signal processing method that extracts signal sources from a composite signal. Since Multispectral images carry Information in the form of spectral signature at different wavelengths, it separates the signal into independent components which are spectrally contrast using higher order statistics. Initially an image is divided into number of patches and each patch is represented as random vector. A mixing matrix is calculated using this vector which separates the image into statistically independent components. By combining first few independent components an image can be formed with 99% of the critical data. However, computational cost is high results seem satisfactory.