New versions of principal component analysis for image enhancement and classification

Two new versions of principal component analysis methods were introduced. The first applies a spatial weighting factor to the samples on the basis of location properties of the samples. The second version involves a new definition of high-order correlation coefficient. The former can enhance the effect of important samples and reduce the influence of the less important samples, whereas the later can enhance the influence of sample high or low values. The principal of the methods were introduced with the case study of identification of Au/Cu associated alteration zones from Landsat TM images.