Noise Reduction and Clutter Suppression in Microwave Imaging and Detection

Abstract

Commercial concealed weapon detection systems are large and expensive and are not suitable to be used as a portable system. Currently, new methods of concealed weapon detection are being developed to build small and compact systems. One such method is based upon the natural resonances of objects; however, no such system has made it to the market due to the low quality of the signals used in the detection algorithms. In this thesis, a prototype concealed weapon detection system is developed and tested for operation in a cluttered environment. This system utilizes the late-time portion of a radar return to extract the resonance information of an unknown target. After proper signal processing and clutter suppression, the signals are classified to determine if the object is a threat. Multiple measurements with frequency-sweep and time-domain systems are used to verify the algorithm. Microwave tissue imaging techniques aim to reconstruct the internal dielectric distribution of the tissue and rely on the dielectric contrast between healthy and malignant tissues. This contrast has been shown to be weak, and therefore, the signals are easily susceptible to noise. This thesis proposes and validates a method for signal-to-noise ratio analysis of complex S-parameter data sets that are used for microwave imaging. A study of de-noising and artifact reduction techniques for microwave holographic imaging is also presented.