Percy (Pei Hei) Lam

  • Digital Roads PhD student Digital Twins, University of Cambridge
Percy (Pei Hei) Lam

Contact

Connect

Location

  • Civil Engineering, University of Cambridge

About

Research Interests

I am interested in exploring how infrastructure can be digitised. Particularly with a massive boom in deep learning and various fronts of artificial intelligence in the past several years, I would like to study on their adoption in infrastructure development, operation and maintenance. With a plethora of data captured on infrastructure by various modalities, improvements on data processing into useful information at scale and visualising information in neat and intuitive manner would contribute significantly to the digitisation.  

Strategic Themes

  • Manual processing of captured visual elements is often the bottleneck in AI. Automation would enable a dramatic increase of the scale of training and inference from new data. 
  • An automated tool to detect defects of general road assets could improve the efficiency of fault reporting and road maintenance. A practical detector also includes the definition of a list of permissible defects on existing road assets. 
  • Working towards the final construction of a digital twin, identified defects would have to be represented in neat and useful manner for subsequent decision-makers to analyse deterioration, make judgements and take action. 

Biography

Percy Lam is currently a PhD student at the EPSRC Centre for Doctoral Training in Future Infrastructure and Built Environment. His PhD research investigates on the construction and maintenance of a digital twin for motorway networks. His academic work is supervised by Dr. Lavindra de Silva, with industrial collaboration supported by Gergely Raccuja (National Highways), Graham Starkey (Costain) and Stuart Hudson (Trimble). 

Percy graduated from the University of Cambridge with MEng and MA (Cantab.) in Civil, Structural and Environmental Engineering in 2016. He embarked on his professional career in geotechnical and construction engineering in a major infrastructure project on the expansion of the Hong Kong International Airport. He was admitted as a Member of the Institution of Civil Engineers and a Chartered Civil Engineer. With an exposure to the intricacy in implementing solutions through his past roles in site supervision, design liaison and quality control, he is driven to carry out research to facilitate the digitisation of infrastructure assets. He is currently researching on relationships of defects on road assets, automatic image annotations and defect detection on pavement and surrounding road assets. These capabilities provide solutions to process data captured over the vast motorway network and will become fundamental building blocks of constructing the UK roads digital twin. 

Publications

Selected publications

Correcting Road Image Annotations Journal article 25 August 2024 International Conference on Computing in Civil and Building Engineering Lam, P., Chen, W., De Silva, L., Brilakis, I Robust Pothole Detection through Dual-pass RGB and Depth Fusion Journal article DIFA: Deep Learning-based Information Fusion and Its Applications Workshop, Affiliated to the 36th British Machine Vision Conference Lam, P., Chen, W., De Silva, L., Brilakis, I.  Analysing the conditions of road assets with a network thinking Journal article 29 March 2023 2023 European Conference on Computing in Construction Lam, P. H., Chen, W., Brilakis, I. CRAAC: Consistency Regularised Active Learning with Automatic Corrections for Real-Life Road Image Annotations Journal article 6 March 2025 2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) Lam, P., Park, S., Chen, W., De Silva, L., Brilakis, I. Toppled Realities: Challenges in Generation and Validation of Synthetic Data Journal article Springer Nature Lam, P., Chen, W., de Silva, L., Brilakis, I.
View full list of publications