A Robust Railway Track Crack Detection System Using Thermal Signatures

In Phase I, Migma Systems propose to develop a robust system for detection of railway track surface cracks. Automated railway inspection is critical for ensuring the safety of public transportation. Rails often break under trains and broken tracks can lead to train derailment. In Phase I, we will make a prototype sensor with low power microprocessor, IR camera, GPS and temperature/humidity sensor. It will be integrated with an existing FRED. Algorithms will be developed to detect track surface cracks through geometric thermal signature. This system also has the self-calibration capability. It does not require external illumination and is suitable for Class 9 train travelling at a speed of 200 mph. In Phase I, we will work with MBTA and install our sensor on FRED of the train for data collection. The collected data will be used for the development of algorithms of track surface crack detection and false alarm mitigation. A data structure will also be formed, which can be used to send event data to the train operators, which including sensor health status. For proof-of-concept, the outcome of Phase I development will be a demo system made of hardware of sensor and software detecting track cracks.