A comparison of the Range-Doppler and Chirp Scaling algorithms with reference to RADARSAT

The Range-Doppler algorithm has long been a popular method for the digital processing of SAR data from spaceborne platforms. The success of this algorithm is due in part to its conceptual simplicity, and in part to a number of approximations which reduce the computational complexity of the method. However, the widespread use of these same approximations has given the Range-Doppler algorithm the (mostly undeserved) reputation of producing inferior image quality, especially with regard to phase. In recent years, the Chirp-Scaling algorithm has been proposed as an alternative to Range-Doppler. The Chirp-Scaling algorithm avoids the complicated and computationally expensive interpolations used by the Range-Doppler algorithm and is more naturally phase preserving. RADARSAT (single beam) data, in view of its moderate squint and multiple PRF Doppler Centroid excursion across an image, is a good test case for both algorithms. Atlantis, with the support of the Canada Centre for Remote Sensing, has developed a desk top software processor which implements both the Range-Doppler and Chirp-Scaling algorithms. In this paper, a comparison is made between the performance of the two algorithms with reference to the RADARSAT sensor. As well, issues of implementation and computational complexity are discussed.