The railway has played a vital role in public transportation and has also been identified as the safest mode of land transport for decades. Key to this is the efficient and constantly improving safety management of some of its major subsystems. Among these, the inspection of railway tracks is an essential part. It ensures that the quality of the rails in operation meets relevant standards and supports appropriate remedial actions to prevent further deterioration or accidents. Rail inspection covers both internal and external defects. External defects, also known as surface defects, are normally derived from long-term rolling contact. They are more common and normally are the point of origin of more severe internal defects. Therefore, the inspection of surface defects is always a significant concern to the industry which should aim to make such inspections more accurate and efficient. In recent years, the development of advanced sensory technologies has automated a lot of rail inspection tasks, which has considerably improved the performance of rail inspection.

By considering the performance of the individual inspection technologies alongside the increased capability of the integrated muti-sensing solution, and through the incorporation of intelligent algorithms that apply complex techniques to the newly available integrated data stream, this thesis demonstrates that an integrated approach provides improved inspection performance.