Submission ID 115323
| Session Title | MO - Driving Change in Maintenance and Operations |
|---|---|
| Title | Standardizing Pavement Condition Assessments Using Spatial and Vision-Based Intelligent Framework |
| Abstract | This study develops a unified pavement condition assessment framework to address the fragmentation in evaluation methodologies across Canadian transportation agencies. Over ten different assessment indices currently in use hinder national-level comparison and resource allocation despite the critical importance of road infrastructure for connectivity. The proposed framework introduces a standardized index based on three key metrics: surface cracking, surface deformation (rutting and potholes), and ride quality (roughness), The methodology integrates advanced LiDAR technology, using a tile-based approach for rutting analysis, an optimized curvature-based algorithm for pothole detection, and computer vision for enhanced crack detection. Preliminary results show high accuracy, with rutting and pothole detection processing speeds of 83 and 88 seconds per kilometer, respectively, and roughness measurements closely matching ground truth data. The proposed standardized combined index showed a strong correlation with separate individual indices for pavement condition assessment. This facilitates the wide adoption of LiDAR and imagery equipment on cities' vehicle fleets for near real-time distress evaluation to make informed maintenance and operation decisions. A first-of-its-kind survey of 14 Canadian transportation agencies informed the framework, identifying technical and institutional barriers to standardization and providing actionable recommendations. The framework's universal index calculation formula ensures compatibility across diverse systems, addressing the urgent need for standardization, particularly given the last national pavement condition assessment conducted in 2004. This research supports data-driven maintenance decisions and optimal resource allocation by enabling accurate, efficient, and consistent evaluation of pavement conditions. The outcomes advance both theoretical methodologies and practical implementation, establishing a foundation for sustainable management of Canada’s transportation infrastructure. |
| Presentation Description (for App) | The study formulates a unified pavement condition assessment framework using spatial-based machine learning and computer-vision to standardize the methods and evaluation of Canada's transportation network. |
| Author and/or Presenter Information | Ali Faisal, University of British Columbia
Suliman Gargoum, University of British Columbia |
