Rational function modeling for spaceborne SAR datasets

Abstract

As a kind of generic sensor model, the rational function model (RFM) has been widely used in geometric processing of optical images, but has not yet been applied to SAR datasets. In this article the feasibility and methodology of rational function (RF) modeling for SAR datasets are investigated. After a review of the mathematic formulation of the RF model and the Range–Doppler model for SAR systems, the feasibility of applying RFM to SAR datasets is analyzed. Afterwards a two-stage approach is proposed as the key technique for SAR RF modeling to solve unknown parameters of RFM in a fast and unbiased way. The effectiveness and advantages of this approach are demonstrated by comparisons with traditional methods. Experimental results obtained for various spaceborne SAR datasets of different processing levels show that RFM is a suitable replacement of the rigorous Range–Doppler model for spaceborne SAR images. Furthermore, the impacts of several factors including the control point grid size, the number of elevation layers, and the orbit precision on SAR RFM solutions are evaluated quantitatively. The results show that the number of elevation layers is a key factor in SAR RF modeling, and its value should be set carefully according to terrain conditions of study areas. Finally, potential applications of SAR RFM are discussed in brief.