Thematic information extraction in high-resolution remote sensing image based on weighted PCA and VBICA

Abstract

The thematic information extraction has been a difficult problem in high-resolution remote sensing application. Principal component analysis (PCA) is able to extract data's independent features on the basis of the second-order statistics, the variational Bayesian independent component analysis (VBICA) not only overcome the inconsistency between the standard ICA model and remote sensing image but also decrease the computational complexity. In view of the characteristics of high-resolution remote sensing, a thematic information extraction method based on weighted PCA and VBICA is presented in this article, and IKONOS high-resolution remote sensing image experiments are performed. The result shows that the classification accuracy of proposed method reaches 78.30% under certain conditions with the suitable number of eigenvectors and weighted values.