Utilization of Spaceborne SAR Data for Mapping

TLDR

Recent advances in automated processing of SEASAT SAR imagery now enable the creation of large‑scale, high‑resolution maps, though standard preprocessing still produces geometrically distorted images. The study evaluates the advantages and limitations of automated SEASAT SAR processing, analyzes associated errors, and presents mapping results. The authors developed unsupervised algorithms that use spacecraft trajectory, imaging geometry, and sensor coherence to perform pixel location, geometric rectification, and mosaicking without ground control points. The technique yields high‑resolution maps, reveals error characteristics, and the study reports results from the SIR‑A experiment and outlines a planned SIR‑B experiment for 1984.

Abstract

Recent developments in automated processing of digital SEASAT SAR imagery have made feasible the generation of large-scale high-resolution maps. Standard preprocessing of raw data into digital images results in geometrically distorted imagery. Computer algorithms have been developed for unsupervised pixel location, geometric rectification, and mosaicking of multiple-image frames without ground control points. These algorithms utilize knowledge of the spacecraft trajectory data, the imaging geometry, and the coherent properties of the sensor to generate the required processing parameters. This paper discusses the advantages as well as the inherent limitations of this technique, analyzes the associated errors, and presents results using SEASAT SAR imagery. Also discussed are the results of the recent shuttle imaging radar (SIR-A) experiment as well as a follow-on experiment (SIR-B) planned for 1984.