Experiences with SRTM/X-SAR phase unwrapping using the minimum cost flow method

Abstract

With the Shuttle Radar Topography Mission (SRTM) flown in February 2000, the most comprehensive interferometric SAR data of the Earth's land mass so far were acquired. At the German Aerospace Center the X-SAR data are processed to a precise and homogeneous Digital Elevation Model (DEM). One of the most critical processing steps is phase unwrapping (PU). We are using an implementation of the Minimum Cost Flow (MCF) algorithm. Now that a total of 16,000 interferograms between /spl plusmn/60/spl deg/ latitude have been processed, this offers the unique opportunity to study the behavior of the MCF algorithm using a very large data pool. In this paper we report on the practical experiences with MCF gained during operational SRTM/X-SAR data processing. We present the results of a throughput analysis. Further, the problems experienced with MCF PU of the X-SAR data are discussed. It is shown how specific adoptions of the MCF cost functions can help to overcome them. The paper also presents the methods of PU quality assessment used in our operational processing.