Tie point generation from overlapping SAR image pairs by Kalman filtering

The paper presents a Kalman filter based approach to determine the heights of a set of tie points seen in a pair of overlapping SAR scenes which have been acquired from different flight paths. The position of registered points as well as flight path information is used to formulate a system of implicit equations which must be solved to determine the heights. The paper shows how Kalman filter techniques may be utilized to solve the outcoming system of equations in a recursive way, not only allowing to solve the system of equations but also to generate error measures in form of covariances. The paper presents the underlying framework, derives the algorithm and presents some first results obtained with airborne data from the DOSAR system