16 Jun Student | Huaxi Huang | Design, Modelling and Simulation
University of Technology Sydney (Supervisor: Jian Zhang)
Rail infrastructure defect detection through video analytics
This project will target the management of defect detection through robust video/image analysis in a clutter environment. Compared with traditional infrastructure maintenance process, inspection through video analysis will significantly improve the efficiency of identifying defects, while also reducing safety concerns by limiting the physical inspection contact between maintenance engineers and infrastructure facilities.
Expected completion date
- Collected, labelled and established a preliminary railway infrastructure defects image dataset.
- Designed an automated image/video railway infrastructure defects recognition framework using computer vision and deep learning technologies.
- Designed three deep learning based image processing algorithms to solve the limited labelled problem in dealing with railway infrastructure defects and fine-grained image recognition.
- Published and submitted several academic papers on leading international conferences and journals about railway infrastructure defects recognition and limited labelled data processing.
Huang, H., Xu, J., Zhang, J., Wu, Q. and Kirsch, C. (2018). Railway Infrastructure Defects Recognition using Fine-grained Deep Convolutional Neural Networks. 2018 Digital Image Computing: Techniques and Applications (DICTA), Canberra, Australia, 2018, pp. 1-8.
Huang, H., Zhang, J., Zhang, J., Wu, Q. and Xu, J. (2019). Compare More Nuanced: Pairwise Alignment Bilinear Network for Few-Shot Fine-Grained Learning. 2019 IEEE International Conference on Multimedia and Expo (ICME), Shanghai, China, 2019, pp. 91-96.