Development of Optimal Thresholding Technique for Shape and Size Detection for Through-the-wall Radar Imaging System

Abstract

Through-the-wall radar (TWR) images of stationary target behind a wall are subject to strong stationary and non-stationary clutter which obscures the target position and size in the image. Stationary clutters are present due to strong reflection from the wall and non-stationary clutter occurs due to multipath, noise, etc. A lot of work has been reported for mitigating stationary clutter successfully for various real scenarios. However, for mitigating non-stationary clutter various authors have reported their work in this field but any concrete result has not been reported so far. Hence, there is a need for an optimal methodology to mitigate non-stationary clutter in TWR images for achieving a high-quality image representing the target position, shape and its size. Till now it is difficult to achieve shape detection of the target from the TWR system. Therefore, in this paper, a novel optimal thresholding technique is proposed to mitigating non-stationary clutter for further enhancement on shape detection of a target using curve fitting and genetic algorithm. The proposed methodology gives a satisfactory result and can prove to be a powerful technique for minimising clutter. © 2019, DESIDOC