Translation and rotational motion compensation techniques for Inverse Synthetic Aperture Radar (ISAR) Imaging of sea vessels

Ebrahim, Azhar

Translation and rotational motion compensation techniques for Inverse Synthetic Aperture Radar (ISAR) Imaging of sea vessels

Thesis / Dissertation

2025

Publisher

University of Cape Town

Department

Department of Electrical Engineering

Faculty

Faculty of Engineering and the Built Environment

Abstract

This dissertation covers range alignment, autofocus, and time window length optimisation for ISAR imaging of sea vessels. The rationale behind the study was to review the range alignment, autofocus, and time window length optimisation techniques found in the literature and to assess the strengths and weaknesses of the techniques. The objective was the development of an ISAR Time Window Selection Processor (ITWSP) to identify focused ISAR images in long radar recordings of sea vessels. This includes the development of a simulator to generate simulated ISAR. A lightweight processor enables efficient scanning of radar recordings to identify focused ISAR images, allowing researchers to prioritise data sets with the highest information content for further analysis without relying on resource-intensive processing. Furthermore, the study was done to determine the performance of range alignment and autofocus algorithms across different datasets. The analysis was done by implementing the techniques on MATLAB, across different measured datasets and simulated datasets. These measured datasets came from maritime vessels, namely the Umoya Omusha vessel and the Zayaan vessel, whereas a simulator was built for the latter. The measures used to quantify the performance of the algorithms were both qualitative and quantitative. The qualitative measure included looking for visual differences in high-range resolution (HRR) profile plots and comparing them. The quantitative measures used were image contrast and entropy. It was found that the sub-integer envelope correlation technique performed better than the integer-shift envelope correlation technique. The best-performing technique, qualitatively and in terms of contrast, was the single reference profile envelope correlation (sub-integer shift) technique. The cumulative averaging reference profile envelope correlation technique achieved the best entropy across the measured datasets. The combination of the weighted multiple scatterer autofocus algorithm and the sub-integer range alignment technique proved to have the most favourable results. Lastly, the Optimal Coherent Processing Interval (CPI) Selection based on Image Contrast Optimisation and the Doppler Spread technique proved to be better than the Maximum Contrast Automatic time window selection technique. Selecting a combination of the best algorithms formed the ITWSP. The alignment algorithm using sub-integer shift range alignment showed better contrast values and overall image quality compared to the integer shift range alignment algorithm, making it the clear choice for processing. For the Umoya vessel, the highest contrast value (19.5903 dB) came from the weighted multiple scatterer autofocus method, which was marginally better than the (19.5643 dB) achieved by the dominant scatterer autofocus method. On the other hand, for the Zayaan vessel, the dominant scatterer autofocus method produced a slightly higher contrast (0.0171 dB) compared to the weighted multiple scatterer method. This highlighted the fact that the choice of autofocus algorithm depends on the dataset. However, the weighted multiple scatterer autofocus method was ultimately preferred for the ITWSP due to its reliable performance across various profiles and its ability to achieve the highest contrast across both datasets. The final ITWSP, which included the envelope correlation sub-integer shift range alignment, weighted multiple scatterer autofocus, and an optimised CPI selection based on contrast and Doppler spread techniques, generated five high-quality ISAR images. From these, the best image was found for both vessels. Although the maximum contrast automatic time window selection also produced five good images, it did not deliver the single best image for either of the vessels.

Keywords

Synthetic Aperture Radar

Sea Vessels

Reference:

Collections

Masters

Full item page