Implementation and Analysis of Autofocus Techniques in ISAR Imaging of Satellites

Sammanfattning

ISAR (Inverse Synthetic Aperture Radar) is an imaging technique that utilizes a stationary radar to image moving targets, providing critical data for target detection, tracking, and identification. However, the inherent target motion often results in defocused images, necessitating the use of autofocus methods. This study implements and evaluates six autofocus methods: two non-parametric and four parametric. These methods were optimized for ideal target data and subsequently evaluated against varying levels of data degradation, with the implementation conducted in MATLAB using synthetic ISAR data. Additionally, two image formation methods were implemented to serve as references. The results demonstrate significant variations in image quality and computational efficiency among the implemented methods. Overall, the non-parametric methods exhibited higher computational speed, while the parametric methods yielded superior image quality, occasionally exceeding the quality of the reference image formation methods. The Range-Centroid Image Contrast-Based Autofocus (RC-ICBA), a parametric method, was identified as providing the optimal trade-off between image quality, computational efficiency, and robustness against data degradation. These findings provide a framework for selecting and optimizing autofocus methods based on specific requirements for quality, efficiency, and robustness.