Identification of the geometrical parameters of cylindrical targets hidden ‎in GPR images using particle swarm optimization (PSO) algorithm

The geophysical ground-penetrating radar (GPR) method is able to detect and identify shallow buried targets ‎especially cylindrical type without any destruction of the medium. GPR method is based on sending the ‎electromagnetic waves generally in the range of 1 MHz up to 1 GHz frequency into the earth and receiving ‎reflected off various buried targets. In present research, this method has been used to identify geometrical ‎parameters of buried cylindrical targets containing burial depth and radius. This task was done through the ‎mathematical relationships between geometrical parameters of cylindrical targets with their GPR hyperbolic ‎response using the particle swarm optimization algorithm (PSO). To achieve this goal first, forward modeling of ‎GPR data by 2-D finite-difference time-domain (FDTD) method using GPRMAX was performed for several ‎synthetic models corresponding to common targets in geotechnical applications and subsurface cylindrical ‎installations. Also in the research, several field surveys were carried out over the well-known buried cylindrical ‎targets in Isfahan University of Technology campus, moreover their GPR response was investigated after ‎employing different processing sequences. In order to extract the geometrical parameters of the buried cylindrical ‎targets in synthetic models and real surveys, the PSO algorithm was used in MATLAB environment. The algorithm ‎performance for 7 produced synthetic models, including cylindrical targets made by different materials, radii and ‎burial depth, as well as the real radargrams of GPR profiles in Isfahan University of Technology campus was ‎verified leading to favorite and reasonable results for the synthetic models and real radargrams, respectively.‎