About the Project
Deep Learning is currently one of the fastest growing fields in Machine Learning and represents and a key area of innovation in computing. The is rapid development of Deep Learning algorithms in combination with GPU technology has meant that that machines can now recognize objects and translate speech in real time. Recently at Glasgow Caledonian we have begun utilising Deep Learning and GPU technology to assist in solving problems in Structural Condition Monitoring. In these projects, Deep Learning based vision systems are used to automatically assess the condition of structural assets using thermal and visible spectral imaging systems. The goal of this project is to extend upon this work to utilize Deep Learning technology in the application of 3D point cloud data obtained using multiview camera systems or LIDAR technology. Specifically the project will investigate classification, segmentation and visualization of 3D data. This will have further applications in structural condition monitoring of assets, and in the development of robotic-based autonomous measurement and analysis applications.
Aims
The aims of this project are:
1. Evaluate current Deep/Machine learning techniques applied to Structural Monitoring.
2. Develop Deep Learning models to tackle specific structural monitoring challenges.
3. Develop novel visualization methodologies for 3D surface reconstruction and abstraction and scene semantic understanding.
4. Develop models capable of exploiting both 2D and 3D input data for complete structural analysis.
Funding Notes
The studentship of £19,300 per year is for a period of three years, subject to satisfactory progress. The studentship covers the payment of tuition fees (currently £4,500 for UK/EU students or £15,000 for international students) plus an annual stipend of £14,800 for UK/EU students or an annual scholarship of £4,300 for international students.
Eligibility: a UK honours degree 2:1 (or equivalent); or a Masters degree in Computer Science, Artificial Intelligence, Industrial Design, Electrical and Electronic Engineering or Mathematics. Equivalent professional qualifications/appropriate research experience may be considered. An IELTS score of 6.5 (or equivalent) with no element below 6.0 is required.
References
1. Mark David Jenkins, Tom Buggy and Gordon Morison, "An imaging system for visual inspection and structural condition monitoring of railway tunnels," 2017 IEEE Workshop on Environmental, Energy, and Structural Monitoring Systems (EESMS), Milan, 2017, pp. 1-6.
2. Simon Stent, Riccardo Gherardi, Björn Stenger, Kenichi Soga, and Roberto Cipolla. "Visual change detection on tunnel linings." Machine Vision and Applications 27, no. 3 (2016): 319-330.
3. A. Boulch, B. L. Saux, and N. Audebert. Unstructured Point Cloud Semantic Labeling Using Deep Segmentation Networks. In Eurographics Workshop on 3D Object Retrieval, 2017.
4. M. M. Bronstein, J. Bruna, Y. LeCun, A. Szlam, and P. Vandergheynst. 2017. Geometric Deep Learning: Going beyond Euclidean data. IEEE Signal Processing Magazine 34, 4 (July 2017), 18–42
Continue with Facebook
