부산대학교 도서관

Antipersonnel mines are hidden weapons that are usually buried close to the surface and are triggered by a foot step of the victim. A sensor that is able to detect minimum metal mines is a ground penetrating synthetic aperture radar (GPSAR). In previous work, an unmanned aerial vehicle (UAV)-based GPSAR was developed and tested for the detection of buried landmines. The real topography was neglected and the interface between air and soil was assumed as a horizontal plane surface. Since the dielectric properties of the soil reduce the propagation velocity of the electromagnetic wave, the interface between air and soil has to be known precisely for subsurface focusing. Neglecting the surface profile has a negative impact on the image quality and prevents the detection of buried objects. In this letter, a novel two-step procedure is used to overcome this issue. In time-division multiplex mode, the radar transmits two frequency-modulated ramps in two different frequency bands. The data of the upper frequency band are used to generate a digital elevation model (DEM) through interferometry. The data of the lower frequency band are used for GPSAR focusing on the basis of the DEM. It is demonstrated that the GPSAR image quality is improved on nonideal planar surfaces and the probability of detection is increased.