| Student: | S.M. Tilon MSc |
|---|---|
| Timeline: | July 2018 - 1 July 2022 |
Road infrastructures are essential to daily life because they connect vital societal assets and thus aid in maintaining the quality of life. For most communities, in the advent of a natural disaster, roads become critical “lifelines” to essential health and rescue operators [4]. Climate change, ageing assets and increased traffic volume and loads, put increasingly more strain on roads effectively accelerating the degradation of roads and consequently negatively affecting the societal quality of life and economic growth.
With accelerated degradation and signs of deferred maintenance, it becomes clear that most western countries face the critical challenge of upkeeping the vital “lifelines” at a sufficient performance level and adapting them to withstand the challenges ahead.
A way of expressing the effect of meteorological- and anthropogenic-induced degradation on road infrastructures is by considering their resilience. Roads are resilient when they can resist and adapt quickly to disrupting events or changing conditions while providing a continuously safe and reliable level of service above a predefined performance threshold. Performance can be expressed as a structure's deterioration or damage [6]. Damages include for example physical deformation, deposition (of foreign material), corrosion and cracks. The level of deterioration of a structure allows road operators to estimate the performance level and remaining service life of a structure by using various analytical techniques, and with these estimations, they can plan and time interventions actions and budget for these actions. Damage can only be reliably measured by utilizing monitoring techniques. Therefore, monitoring is vital in understanding an infrastructure’s resiliency because it aids to determine the correct performance level [3]. However, monitoring road infrastructures is becoming increasingly complex for the reasons addressed above and with the increased demand to intensify monitoring at scale.
Current monitoring practices are inadequately adept to monitor road infrastructures under these circumstances. Traditional road infrastructure monitoring consists of human visual inspections [2,8]. The main advantage of this practice is the ease of integration of expert knowledge and the flexibility in deployment. The main downsides are related to subjectiveness, ambiguity and speed. Modern inspection techniques, such as fixed sensors techniques, are increasingly used to support human inspections. Remote sensing technology, using active or passive, and optical or multispectral sensors, attached to a ground-based or airborne platform [1,5,9], is used to inspect infrastructure from a distance at larger scales. Yet, one concern remains. Modern monitoring technology is insufficiently capable of holistic monitoring, yielding infrastructures unprepared for climate change. Infrastructures physically reside in a larger context of natural and anthropogenic elements. Yet, they are often regarded as stand-alone objects within monitoring practices. Other (natural) processes occurring within the road corridor, the area encompassing the main road infrastructure and its surroundings, are often insufficiently considered.
In line with the challenges facing road infrastructures and the identified shortcomings of existing monitoring practices as laid out above, this dissertation deals with the following research topics:
- Emerging and state-of-the-art machine learning techniques to identify road infrastructure damages. In more detail, this research direction explored the usage of:
- Generative Adversarial Frameworks (GAN) for damage identification of buildings and road surfaces in a training-data-scarce environment.
- Multi-task deep learning for joint, real-time and edge-embedded inference.
- Platforms and optical imagery to identify road infrastructure damages in a post-hazard and routine scenario. Here the usability of satellite, UAV and ground-based vehicle platforms was explored in their ability to detect damages.
- Hardware-accelerated UAVs for optimized real-time damage and anomaly identification. Here a hybrid UAV (the DeltaQuad), a fixed-wing vertical take-off and landing (VTOL) UAV was customized with an onboard computation device and a novel communication and analytical processing framework to enable real-time road corridor inspections.
- Semantic reasoning for qualitative and informed decision-making to match the information needs of road operators.
This dissertation has been conducted with the support of and in collaboration with the PANOPTIS project, funded under the European Union’s Horizon 2020 research and innovation programme. PANOPTIS aimed at the ”Development of a Decision Support Framework for Increasing the Resilience of Transportation Infrastructure Based on Combined Use of Terrestrial and Airborne Sensors and Advanced Modelling Tools”[[1] panoptis.eu].
Figure 1. Data acquisition at the PANOPTIS pilot site with the fixed-wing VTOL (the DeltaQuad).
Figure 2. Example of output produced by a vehicle detector. From [7].
- L. Bruzzone, A. Haghighattalab, A. Mohammadzadeh, M.J. Valadan Zoej, M. Taleai, Post-earthquake road damage assessment using region-based algorithms from high-resolution satellite images, Image and Signal Processing for Remote Sensing XVI, 2010.
- K. Chen, M. Lu, G. Tan, J. Wu, CRSM: Crowdsourcing Based Road Surface Monitoring, 2013 IEEE 10th International Conference on High Performance Computing and Communications & 2013 IEEE International Conference on Embedded and Ubiquitous Computing, 2013, pp. 2151-2158.
- D.M. Frangopol, Life-cycle performance, management, and optimisation of structural systems under uncertainty: accomplishments and challenges1, Structure and Infrastructure Engineering 7 (6) (2011) 389-413.
- S. Hallegatte, J. Rentschler, J. Rozenberg, Lifelines : The Resilient Infrastructure Opportunity, World Bank, Washington, DC, 2019.
- H. Ma, N. Lu, L. Ge, Q. Li, X. You, X. Li, Automatic road damage detection using high-resolution satellite images and road maps, 2013 IEEE International Geoscience and Remote Sensing Symposium-IGARSS, IEEE, 2013, pp. 3718-3721.
- K.C. Sinha, S. Labi, B.R.D.K. Agbelie, Transportation infrastructure asset management in the new millennium: continuing issues, and emerging challenges and opportunities, Transportmetrica A: Transport Science 13 (7) (2017) 591-606.
- S. Tilon, F. Nex, G. Vosselman, I. Sevilla de la Llave, N. Kerle, Towards Improved Unmanned Aerial Vehicle Edge Intelligence: A Road Infrastructure Monitoring Case Study, Remote Sensing 14 (16) (2022).
- C. Zhang, A. Elaksher, An Unmanned Aerial Vehicle-Based Imaging System for 3D Measurement of Unpaved Road Surface Distresses, Computer-Aided Civil and Infrastructure Engineering 27 (2) (2012) 118-129.
- D. Zhang, Q. Li, Y. Chen, M. Cao, L. He, B. Zhang, An efficient and reliable coarse-to-fine approach for asphalt pavement crack detection, Image and Vision Computing 57 (2017) 130-146.
Meet the team
