We brought forward a hyperspectral imagery unmixing methods based on the posterior probability of relevance vector machine.In the hyperspectral imagery unmixing methods based on the posterior probability of support vector machine

the posterior probability have to be got by sigmoid function approximately

and the model need to set regularization coefficient by cross-validation.Relevance vector machine is a sparse model in the Bayesian framework

its mathematics model doesn’t have regularization coefficient and its kernel functions don’t need to satisfy Mercer's condition.In this paper

we firstly analyzed the disadvantages of the support vector machine for hyperspectral imagery unmixing

introduced the sparse Bayesian model and the model parameters inference

and selected the fast sequential sparse Bayesian learning algorithm.Through the unmixing experiments of PHI imagery

the advantages of the relevance machine for hyperspectral imagery unmixing were given out.