Train-mounted GPR for high-speed rail trackbed inspection

Recent tragic rail accidents have added momentum to the international interest in methods to automate the inspection of railway trackbed (ballast). Presently, much of the trackbed inspection is carried out visually by railway staff, with results often unreliable and inconsistent, let alone slow and expensive. It is envisaged that GPR can be used to provide rapid, objective and quantitative information about the depth and degree of deterioration of the ballast with minimal disturbance to the actual trackbed. This research seeks to automate the processing and interpretation of such data to the extent whereby on-site interpretations may be achieved with minimal expert intervention. This is through developing new image and signal processing tools specifically for GPR data and the range of anomalies found on the trackbed. The most efficient way to monitor condition of track and other infrastructure assets is by means of a train, which can collect data for many parameters simultaneously, where possible at normal line speed. A multichannel ground- penetrating radar system is presented, capable of operating at speeds of up to 200 kmph. A road-rail variant of the system can collect up to 6 simultaneous continuous channels across the track, and deliver on-site interpretation of ballast thickness and quality, irregularities, weak spots and utilities. This is achieved by the application of novel multivariate signal and image processing techniques to automatically detect, quantify and map variations in ballast depth and condition. Multi-resolution texture analysis techniques are applied to enable automatic characterisation and classification of regions of interest within the radargrams. The system can probe the ballast both underneath and between the sleepers, thus identifying potential problems with individual sleepers. A quantifiable measure of ballast and formation condition is provided that can be used to QC new renewals or provide an "inventory" of ballast condition. This integrated inspection system offers significant productivity and reliability improvements over conventional GPR techniques. By combining information from track geometry surveys, GPR, optical and infra-red video and other co-located datasets, predictability of track degradation is improved, while reducing the required track possession time and providing on-site turnaround of site interpretation, costs are reduced significantly.