Spatial resolution enhancement of rotational-radar subsurface datasets using combined processing method

Abstract

Abstract Effective visualisation of railway tunnel subsurface features (e.g. voids, utilities) provides critical insight into structural health and underpins planning of essential targeted predictive maintenance. Subsurface visualisation here utilises a rotating ground penetrating radar antenna system for 360° point cloud data capture. This technology has been constructed by our industry partner Railview Ltd, and requires the development of complimentary signal processing algorithms to improve feature localisation. The main novelty of this work is extension of Shrestha and Arai’s Combined Processing Method (CPM) to 360° Ground Penetrating Radar (360GPR) datasets, for first-time application in the context of railway tunnel structural health inspection. Initial experimental acquisition of a sample rotational transect for CPM enhancement is achieved by scanning a test section of tunnel sidewall - featuring predefined target geometry - with the rotating antenna. Next, frequency data separately undergo Inverse Fast Fourier Transform (IFFT) and Multiple Signal Classification (MUSIC) processing to recover temporal responses. Numerical implementation steps are explicitly provided for both MUSIC and two associated spatial smoothing algorithms, addressing an identified information deficit in the field. Described IFFT amplitude is combined with high spatial resolution of MUSIC via the CPM relation. Finally, temporal responses are compared qualitatively and quantitatively, evidencing the significant enhancement capabilities of CPM.