top of page

Fighting potholes with AI


The last few years has seen an explosion of research projects, trials and new products which aim to spot potholes and other road defects automatically using machine learning, a type of artificial intelligence (AI).

A machine learning algorithm is a bit like a toddler (albeit a very slow learning one): you show it lots of different images of the same object, tell it what it is, and eventually it can recognise the object by itself. So, all the trials involve a means of collecting images combined with software that then processes those images and tells you about the defects (very much faster than a toddler could do or, indeed, a human of any age).

Here are some of the councils that are putting AI to the test:

Kent County Council, with Amey, is trialling a system which combines cameras, sensors and AI to collect information on potholes, cracks and also assets such as signs, white lines and street furniture. Funded through the Government’s ADEPT programme, the system can use existing cameras on buses and council vehicles and is also trialling stereo cameras which can give a 3D perspective.

Route Reports, the company behind the technology, was set in 2017 when CEO Connell McLaughlin was a teenager. The company has also worked with a major train operator in the US, using vibration data to detect defects on tracks.

West Lothian Council is working with Finnish technology company Vaisala. The £14,000 AI trial has been funded from the council’s Digital Capital Fund.

Vaisala, which is also involved in the Mars Rover Perseverance programme, has a tool called RoadAI which combines geospatial video, taken using any vehicle, with AI that can spot multiple types of defect, providing the video and information through a web portal. RoadAI has also been used on rural roads in Suffolk and in Bexley in outer London.

Durham County and Blaenau Gwent Councils, funded by GovTech Catalyst, are both looking at the viability of a system from construction software company GPC Systems which deploys 3D cameras on a variety of council vehicles.

GPC’s Highway Measure software uses 3D images to analyse the width and depth of the hole together with the volume of replacement material required. GPC has recently signed an agreement with asset management software firm Yotta to integrate its 3D Construction tool with Yotta’s Alloy asset management software.

Looking into the future…

Researchers at the University of Liverpool are trying to go one step further, developing a robot which can both detect and repair cracks before they become potholes. Having developed the technology over four years, the two scientists involved have set up a new company Robotiz3d, created in partnership with electronic design company A2e.

Thermal Road Repairs is a green technology company which supplies systems to improve the quality, cost and time efficiency of road repairs and paving – at a far lower environmental cost than traditional methods. We invest significantly in R&D, to create new technologies and to continuously improve our existing ones.

High output. Low emission. Permanent solution.

7 views0 comments

Recent Posts

See All
bottom of page