Analysis of gpr data through interpretation of pre-processed images obtained by a multi-agent approach to identify pipes in water supply systems

This work focuses on the development of easy application procedures for visualizing the characteristics of the components of water supply systems (WSS), quickly and by non highly qualified staff. We study databases related to the underground obtained with GPR (ground penetrating radar). In this study we perform GPR imaging of pipes of four different materials commonly used in WSSs buried in dry soil. The data obtained from the survey are pre-processed with a multi-agent method consisting in a race of agents. Subsequently, analysis and interpretation of the results, seeking to generate forms that allow a quick understanding, are performed. These forms are analyzed in order to assess the feasibility of pattern recognition revealing existence of pipes. The results are promising in the attempt of generating suitable databases and parameters to train intelligent systems for characterizing components of WSS. © 2013 The authors and IOS Press. All rights reserved.SCOPUS: cp.k