Technology being developed at Nottingham Trent University aims to detect the early signs of potholes and determine their severity.
The technology, which is being developed by a team led by research fellow Dr Senthan Mathavan, scans roads for ravelling, a situation where the loss of aggregates from asphalt leads to potholes and cracks.
Combined with 2D and 3D scanners on a pavement-monitoring vehicle, a computer vision algorithm can examine the road with accuracy at traffic speed during the day or at night.
The system, which can be installed into existing hardware, works by detecting different textures of the road to identify ravelling and distinguishes it from shadows and blemishes such as tire marks, oil spills and recent pothole repairs.
‘The existing hardware remains untouched,’ Dr Mathavan told The Engineer. ‘Our technology makes use of the data from this hardware and uses a novel algorithm to look for the signs of ravelling.’
During their research, the team found that the technology detected road surfaces correctly in all 900 images tested. It took approximately 0.65 seconds to 3D process the ravelling measurements, but it is believed that this could be reduced further.
Next steps include benchmarking of the technology with respect to conventional, non-automated methods for measuring ravelling.
‘In addition, plans are also afoot to compare the detection algorithm’s performance against human inspectors,’ said Dr Mathavan. ‘Based on this benchmarking, we will fine tune knobs in the software. In addition, repeatability of the results needs to be established through an extensive series of tests.’
The research was published in Transportation Research Record and also involves Dr Mujib Rahman of Brunel University, Martyn Stonecliffe-Jones of Dynatest UK Ltd, and Dr Khurram Kamal of the National University of Sciences and Technology in Pakistan.
Potholes cost motorists approximately £2.8bn a year and local authorities currently pay out over £30m to cover compensation claims related to them. Despite this, few technologies exist to detect potholes propagating.
‘There is one known software to detect ravelling,’ said Dr Mathavan. ‘However, the workings of it are not transparent and reservations exist on its usage in the road monitoring industry – the reason why Dynatest wanted to do this research in the first place. However, when it comes to research into ravelling detection there were no previous studies and ours is the first effort.
‘For a difficult-to-detect defect such as ravelling, we first need extensive research in academia before reaching a wider consensus on what works and what does not. This process, although had been followed for pothole and crack detection techniques, has not been adopted for ravelling. We believe [that] our paper in this area kick-starts that process for ravelling detection. There is still much to be done.’
Dr Mathavan added that a number of local councils have expressed an interest in deploying and testing the effectiveness of the technology.