3D rail modelling and measurement for rail profile condition assessment

To ensure safety and reliability, compliance to rail regulations and efficient predictive maintenance, regular condition assessment needs to be carried out on the railway infrastructure. The consequence of not carrying out condition assessment measurement results in poor knowledge of track conditions that should trigger maintenance and rehabilitation interventions. With the introduction of higher speeds, safety requirements for freight and passenger trains, track occupation limitation, cost and productivity improvements in maintenance requirements, more efficient condition assessment and maintenance systems are required. Rail profile measurements and visual inspection of rails are some of the fundamental condition assessments requirements for determining the wear rate and surface defects on rails. This paper proposes the use of 3D image acquisition and modelling to extract the condition of rails (through rail profile measurements) and detection of surface defects. This is supplemented by manual measurement methods to validate and confirm the accuracy and reproducibility of the vision based approach. Various image processing techniques are reviewed, and a selection criterion taking into consideration external influences such a variation in lighting, vibrations on the railway track environment, amongst other conditions are discussed.