Use of a sub-array statistical approach for the detection of a buried object

A hybrid electromagnetic-statistic approach for the detection and localization of a perfectly-conducting circular cylinder, buried in a lossless half-space, is presented. We use the results of a cylindrical wave approach forward solver as input data for our detection procedure. We use a sub-array processing structure and apply several algorithms for the direction of arrival estimation. By triangulating the found directions of arrival, a set of crossings, condensed around the object locations, is obtained. To process the crossing pattern, we developed a statistical model for the crossings distribution and employed hypothesis testing procedures to identify a collection of small windows containing the target. By defining a suitable threshold from a desired false alarm rate and dividing the region in small windows it is possible to ascribe each window to the ground or to the object. Numerical results are presented for a cylinder in a vacuum and in a dielectric half-space, both in a central and in a peripheral position with respect to the array centre. Different values of the cylinder radius and of the distance from the array are considered.