Model Pyramid Based Decomposition and Fast Browsing Scheme for Magnanimity Imagery

Images of remote-sensor (RS) become to be an important way on Earth observation. However, the high resolution of most RS images causes that the images contain magnanimity data sets, which makes it difficult to browse using usual schemes. In this paper, a fast browsing scheme based on pyramid decomposition is presented. Unavoidably losing information after pyramid decomposition, it is expected not only to observe the global characters but also to reserve the local characters. In this paper, several pyramid decompositions are presented, such as linear Haar wavelet etc. Then a new pyramid decomposition based on AR model is proposed, called model pyramid decomposition (MPD). An evaluation to the results from schemes above is given at the end of the paper