Aiming at the detector filling rate not equaling to 1 and there existing contraposition bias in actual micro-scanning system, the corresponding sub-pixel thermal imaging algorithm was investigated.

When applying this method for the trehalose samples, the relative error is not more than 3.

Compared with soil CRM certificate,a better determination result can be obtained with relative error being between 0 ~ 10% except sodium oxide and magnesium oxide.

According to the 1/2 standard of American Clinical Laboratory Improvement Amendment (CLIA,88),the relative bias of the reference analyzer and other assessed analyzers were determined.

At the same time,the paper gives an illustration of this new model and compares the relative deviation between EIV prediction model and Multivariate Regression Prediction Model,and shows that the EIV prediction has better precision of prediction.

This paper tries to find out the changing regularity of relative deviation by calculating the settlement quantity of foundation soil according to two different choosing value standards and comparing them with each other.

Through establishment of the mechanical model of the shaft support the variation of the bearing static load produced by the shaft alignment deviation is calculated and further the related dynamic characters of the bearing is also calculated to evaluate the stability of shaft running.

Concerning the biases occurred in sensors measuring data (or range,azimuth),this article gives an estimation of multisensor relative divergence based on Kalman filter technique, by selecting a master sensor as comparison criterion.