Improving Segmented Interferometric Synthetic Aperture Radar Processing Using Presumming

Within the last decade, Interferometric Synthetic Aperture Radar (INSAR) processing has emerged as an important extension of conventional SAR because it can be used to infer topographic heights of the imaged terrain. The demand for high-resolution, large coverage-area topographic data is driving new INSAR systems to acquire data in long strips. When processing long strips of INSAR data, the motion of the sensor requires processing the data in patches, each with its own set of motion parameters. Conventional techniques for aligning these patches are not adequate for INSAR applications, where systematic height errors of roughly 2 m are visible in the topographic images. A data-smoothing procedure known as presumming can reduce these errors by rejecting aliased frequencies in the azimuth spectrum. Presumming operations were implemented in an existing INSAR processor to reduce the patch boundaries in the topographic images. It was found that the location of the presumming operations is critical to reducing patch boundaries effectively. Presumming after motion compensation smoothes features in the azimuth direction, thus making patch boundaries less distinct, but it does not specifically reduce the patch boundary errors. Implementing presumming prior to motion compensation operations may improve results.