Robust cascaded frequency filters to recognize rebar in GPR data with complex signal interference

Abstract Automatically recognizing rebar by using ground penetrating radar (GPR) is an ongoing research area that attracts much attention from researchers and practitioners. Existing studies in the literature have presented various methods to recognize rebar from GPR data, but these methods are mainly focused on straight-forward cases and do not systematically address the signal interference challenges. In this manuscript, the authors propose a novel method to address these challenges based on designing robust cascaded frequency filters. The frequency filters are pertinently designed to decompose GPR data into different components according to the directional transformation of the GPR image, which are then reconstructed as a refined hyperbolic pattern based on a similarity comparison between right and left tail images. Finally, the proposed method is validated using both synthetic and actual on-site datasets and the obtained results are compared with the results of two other state-of-the-art methods existing in the literature.