Subsurface Object Recognition in a Soil Using UWB Irradiation by Butterfly Antenna

The problem of recognition of a cylindrical perfectly conducting object buried in a soil is solving by artificial neural network that takes data from subsurface radar with the use of impulse electromagnetic wave. The air-ground interface is irradiated by the butterfly antenna excited with ultrawideband short impulse current. Electromagnetic problem is solved by FDTD simulation. The results of the calculation in form of time dependences of differently polarized electrical components of received fields in four points are used for artificial neural network training for determination of the cylinder placement. The recognition abilities of the network are verified by noisy input data with different levels of additive Gaussian noise and received for shifted positions of receiver.