← Back to Projects

AI-Remote Inspection

Overview

This project focuses on developing AI-powered solutions for remote inspection of industrial assets and infrastructure. Using computer vision and deep learning, we enable automated defect detection, anomaly identification, and condition monitoring without requiring physical presence at inspection sites. This work included several funding bodies and industrial partners as shown below.

Research Areas

Our work encompasses advanced image analysis and multi-modal sensor fusion techniques. We develop robust AI models capable of detecting structural defects, corrosion, cracks, and other critical issues across a wide range of industrial sectors. In addition, we have developed and tested methods for inspecting underwater pipelines. Virtual reality is employed to deliver an immersive experience for domain-expert engineers, ensuring they remain actively involved and informed throughout the inspection and decision-making process.

Applications

The developed methods were developed mainly for Oil and Gas industry, but can also be applied to other industries including:

  • Oil and gas pipeline inspection
  • Wind turbine blade assessment
  • Bridge and infrastructure monitoring
  • Power line inspection
  • Industrial facility maintenance

Impact

The remote inspection systems we’ve developed have reduced inspection costs, while improving detection accuracy and inspector safety.

Funding

In total, we received £600,000 in funding from Innovate UK, AISUS Offshore Limited, and the NetZero Technology Centre, to support our research on Remote and Pipeline Inspection using AI. It also benefited from the support of industrial partners—BP, TotalEnergies, and Chevron—who contributed domain expertise and pipeline inspection data.

Selected Publications

[1] U. Gurung, M. S. Mekala, C. F. Moreno-Garcia, N. Zolnhofer, B. Marshall and E. Elyan, “Offshore Asset Inspection Redefined: Expert- Validated Deep Learning for Critical Defects,” 2025 IEEE 12th International Conference on Data Science and Advanced Analytics (DSAA), Birmingham, United Kingdom, 2025, pp. 1-10, doi:10.1109/DSAA65442.2025.11248009

[2] T. Dang, T. T. Nguyen, A. W. C. Liew, and E. Elyan, “Event classification on subsea pipeline inspection data using an ensemble of deep learning classifiers,” Cognitive Computation, vol. 17, p. 10, 2025, doi: 10.1007/s12559-024-10377-y.

[3] H. Farhadi Tolie, J. Ren, and E. Elyan, “DICAM: Deep inception and channel-wise attention modules for underwater image enhancement,” Neurocomputing, vol. 584, 2024, doi:10.1016/j.neucom.2024.127585.ucom.2024.127585

Team

The team includes two PhD students, 1 postdoctoral researcher, two research assistants (KTP Associates) and three academics.