The use of estimation techniques to reduce noncoherent polarimetric measurement errors

Polarimetric measurements contain errors owing to random thermal energy in the receiver and system phase noise. These errors can be reduced using estimation techniques that account for general properties of the Mueller matrix. In this paper, we show how Kalman (1960) filter techniques improve and speed up the estimation of a target's Mueller matrix when the noncoherent measurement technique is used. Measurements obtained with a 95-GHz polarimeter demonstrate that fewer independent samples of the backscattered fields are needed to obtain an accurate estimate if Kalman filtering is used than if a pseudo inversion process is used.<
>