Kalman Filter for Removal of Scalloping and Inter-scan Banding in Scansar Images

Spaceborne synthetic aperture radar (SAR) plays more and more important role in Earth observation science, especially with ScanSAR mode which provides wide-swath coverage with moderate resolution. However, scalloping and inter-scan banding (ISB) are two major artifacts, which significantly degrade the quality of ScanSAR images. In this paper, a novel technique for removal of scalloping and ISB in ScanSAR images is proposed. Scalloping and ISB artifacts are modeled by two-dimensional gain and offset parameters varying as function of both azimuth time and range position. The gain and offset parameters can be split into azimuth and range components. The variations of gain/offset with respect to azimuth and range positions would represent scalloping and ISB artifacts respectively. In the proposed technique, recursive and minimum mean square error (MMSE) estimates of azimuth gain/offset parameters are found out by using Kalman filter for each azimuth location in a subswath by considering corresponding range samples as observation vector. Subsequently, range gain/offset parameters causing ISB artifacts are estimated by using Kalman filter for each range positions by considering azimuth samples as observation vector. The MMSE estimates of gain/offset parameters are used to directly remove scalloping and ISB artifacts. The proposed scheme was applied on simulated as well as calibrated real ScanSAR images. The experimental results exhibited the potential of proposed technique to be used as post processing tool for enhancing ScanSAR image quality. Received 21 August 2012, Accepted 27 September 2012, Scheduled 5 October 2012 * Corresponding author: Jie Chen ([email protected]).