A Ground Penetrating Radar (GPR) survey for subsurface investigation yields a large amount of data; within this data only a small percentage represent the actual targets of interest. The current GPR data processing for underground utilities relies on skilled operators and involves many computationally expensive and time consuming processing stages before any information about the subsurface is concluded. These results are often based on broad assumptions and involve human intervention, which is a source of inconsistency and error, hence rendering the results to be unreliable. The data...