2d Atmospheric Artifact Compensation with Multiple Regression Model in Ground-based Sar over Mountainous Areas

This paper proposes a new technique for Atmospheric Phase Screen (APS) compensation based on a 2D Multiple Regression Model (height-dependent) in Ground-Based (GB) zero-baseline synthetic aperture radar (SAR) acquisitions over mountainous areas with steep topography. For this purpose, polarimetric measurements acquired using the GB-SAR sensor developed at the Universitat Politècnica de Catalunya (RISKSAR) are employed. This technique has been applied in the mountainous environment of Canillo in Andorra where there is an active landslide that nowadays is being reactivated coinciding with strong rains. Data sets are being acquired at X-band during one-year measuring campaign (October 2010 – October 2011). The effects of the atmosphere variations between successive campaign acquisitions and the way to compensate it are treated here in detail. The need to compensate for the resulting phase-difference errors when retrieving interferometric information is put forward. A new compensation technique is then proposed and evaluated using the control points placed inside the observed scene atmospheric artifact compensation in presence of steep topography.