Home > MEng Dissertation: Richard van Schalkwyk

MEng Dissertation: Richard van Schalkwyk


Richard van Schalkwyk

Click here to read the interview

Richard van Schalkwyk grew up in Pretoria, and completed his undergraduate degree at the University of Pretoria with distinction in 2009.

From 2011 to 2014, he attended various modules of the MEng (Radar and Electronic Defence) at the University of Cape Town, graduating in 2014 with distinction.

His supervisor was Prof Mike Inggs and the topic of his MEng dissertation was “Scattering Centre Extraction in High Resolution Radar Imaging”.

He is currently working at the CSIR DPSS in the field of Radar.

Click on the photo to read an interview with him, which is part of the ‘Meet our Alumni‘ series.



Van Schalkwyk, Richard F. Scattering Centre Extraction in High Resolution Radar Imaging. MEng (specialising in Radar and Electronic Defence) Dissertation. Department of Electrical Engineering, University of Cape Town, 2014.



The objective of this dissertation is to establish a preferred approach to decomposing electromagnetic scattered field data into a sparse representation of scattering centres. In order to complete the research objective, a performance comparison between two scattering centre extraction techniques is presented. The selected techniques were chosen, as preferred candidates, from two distinct classes of algorithms. The first technique was chosen from a class of deconvolution-based algorithms and the second from a class of super-resolution techniques.

The performance comparison places emphasis on applications in controlled environments such as high fidelity electromagnetic simulations which allow for the generation of high-quality high signal-to-noise ratio data. The study makes use of synthetic scattered field data generated in Matlab and FEKO (a computational electromagnetic software package) and evaluates the performance of the techniques in terms of: resolution, measurement precision and radar cross section re-synthesis. The results from the evaluation have shown that under low signal-to-noise ratio conditions, the deconvolution technique performs better; while under high signal-to-noise conditions, the super-resolution technique is shown to perform better.

A performance-limiting characteristic, in the use of conventional polar reformatting in the super-resolution technique, was identified during the study. This led to the development of an iterative polar reformatting technique which makes direct use of estimated scattering centers. Initial results have shown that the new approach offers significant performance improvement and highlighted that attributes from both classes add value in the preferred approach.