The underground cavity detection within invisible medium analytically using the subsurface radar is extremely difficult because of inhomogeneous property of the surrounding medium. However, when the statistical characteristics of the radar signal is known, the method of the cavity detection is able to construct using the proper signal processing. The cavity detection is made possible by using the statistical variance of radar signal for reasons of it implies the wave scattering degree. So we make easily to evaluate the position of the underground cavity.
In this paper, it is described that the amplitude distribution of the subsurface radar signal obeys to the Weibull distribution and the statistical variance is calculated by Log/CFAR processing.
The suitable choice of data cell size at the sampled radar signal is experimentally considered. Finally, We showed the variance of field test and the relations between the variance and the cavity presence were discussed.