In this paper

a particular review on traditional remote sensing statistical analysis model is made and their defects are analyzed from several aspects especially brought by the statistical assumption of its unimodal probability density functions and parametrical density distribution model. On the basis of traditonal statistical model

some intelligent ways of extensions are proposed

where the major attentions are paid on integration with geo knowledge processing unit and computation with other artificial intelligence approaches such as fuzzy logic

conventional geo analysis

robust statistics

neural computation

and evolutionary computation. Finally

several detailed approaches of extensions of statistical remote sensing analysis model

such as RBF

GMDD

and knowledge integrated model

are introduced and discussed briefly.