Feature registration of large resolution difference non-homologous SAR image pairs for sea ice drift tracking

Abstract

At present, SAR images from same source are widely used in the field of sea ice drift tracking. Due to the longer revisit time of homologous spaceborne satellites, only an average velocity can be determined. For longer time intervals, velocities due to short-duration events such as storms are lost. Synthetic Aperture Radar (SAR) images from different sources make it easy to construct image sequences with short time intervals. However, the resolution and noise level between non-homologous SAR image pairs often differ greatly. When there is a relatively large resolution difference between image pairs, the areal features between image pairs are very different, which increases the difficulty of feature registration. In this paper, a super-resolution reconstruction method is proposed to solve the problem of resolution difference between image pairs for sea ice drift. This method can significantly improve the quality of feature registration of image pairs from different SAR sensors. We demonstrate through several examples the effectiveness of the method in feature matching of large resolution difference images from different SAR sensors.