Vision Based Railway Track Monitoring using Deep Learning
This work addresses the challenge of automated railway track monitoring for maintenance and safety, representing an incremental improvement by applying transfer learning to a domain-specific problem.
The paper tackles the problem of automating railway track defect detection by using transfer learning to train deep learning classifiers that generalize to real-world scenarios with limited labeled data, achieving efficient detection of defects and assets across diverse conditions.
Computer vision based methods have been explored in the past for detection of railway track defects, but full automation has always been a challenge because both traditional image processing methods and deep learning classifiers trained from scratch fail to generalize that well to infinite novel scenarios seen in the real world, given limited amount of labeled data. Advancements have been made recently to make machine learning models utilize knowledge from a different but related domain. In this paper, we show that even though similar domain data is not available, transfer learning provides the model understanding of other real world objects and enables training production scale deep learning classifiers for uncontrolled real world data. Our models efficiently detect both track defects like sunkinks, loose ballast and railway assets like switches and signals. Models were validated with hours of track videos recorded in different continents resulting in different weather conditions, different ambience and surroundings. A track health index concept has also been proposed to monitor complete rail network.