Clutter reduction and object detection in surface penetrating radar

Abstract

This paper addresses the problem of detecting buried objects using a surface penetrating impulse radar. A big problem is that reflections from the ground surface will obscure the often much weaker target signal. The ground reflections and the target signal are modeled using a very simple model. This model is used to derive a detector based on a likelihood ratio test, where unknown model parameters are estimated using a maximum likelihood approach. Even though the model is very simple, the detector is found to perform well. The detector is evaluated using real data from an impulse radar system. Both metallic and non-metallic buried objects can successfully be detected.