Calibrating and mosaicking surface velocity measurements from interferometric SAR data with a simultaneous least-squares adjustment approach

摘要

The repeat-pass interferometric SAR (InSAR) technique has been well established as a precise means to extract two-dimensional surface velocity fields. Many applications require velocity measurements over a large area and multiple InSAR image frames from the same and/or adjacent orbits are needed to achieve full ground coverage. The conventional frame-by-frame processing approach often causes velocity discrepancies and discontinuities between adjacent frames. In addition, absolute velocity information may not be derived for some image frames due to a lack of velocity control points. We present a unified simultaneous least-squares adjustment method for calibrating and merging surface displacements derived from multiple InSAR image frames using the speckle tracking method. Observation equations have been mathematically derived for velocity control points and tie points. The major advantages of our method include consistent and smooth transition of velocity measurements between adjacent frames and dramatic reduction of velocity control point requirements. For those frames devoid of velocity control points, absolute velocity measurements can still be derived with this method based on distant velocity control points in other frames. The benefits of our method are demonstrated and evaluated by using Radarsat InSAR data for Antarctica acquired in 1997.