This paper first introduces the theory of Beyesian Kriging briefly, and analyses the problems when the parameters of conditional variogram and variogram of soft data are estimated, then basing upon the theory of Beyesian Kriging, the interpolation method of Beyesian residual Kriging is induced.

The main characteristic of the method is that it introduces a prior information matrix (Cm) of model parameters,and fastens the convergence of the iteration process.

This study centers on utilizing a prior information to improve the resolution of integer ambiguity.

Method for obtaining prior information of electrical impedance tomography;

Simulation shows that this method is easily implemented and can objectively reflect the reliability degree of each sensor without knowing any prior information.

The algorithm smooths the gray histogram of the test image with Harr wavelet to remove the "pseudo valley",and fixes a range of the threshold according to the prior information in the standard image,and then finds out the valley of the histogram in the range and that is the threshold.

The iteration formula and determination of iteration step are set up according to a priori information of rock and concrete.

The problem of using priori information to cract classical confidence limit is discussed in this paper.

In order to improve the rationality during the modeling,we proposed the semi-parametric regression model constrained with priori information.