Using the decrease of the information entropy of a measurement and the prediction and updating matrixes of covariance in Kalman filter equations
we deduced the information gain in a measurement for target detecting or tracking.And if given the maximum allowable covariance matrix P
the maximum time interval n can be obtained.It is useful for the multi-sensor target tracking and even for the following target attributes information fusion
the target recognition and the form of tactical situation.The results have offered an effective approach for sensor resources assignment.