Case study - Road safety

Domain:  Infrastructure

Client: Amey / Highways England

Other meta data about project:

Highways England currently inspect the strategic road network for pot hole damage by having two people in a car, driving every mile of the network each week – a driver and an observer.


Under Covid-19 regulations, it was no longer feasible to have two people in the same car. The driver clearly cannot observe and drive, so dashboard camera film was used instead, with someone watching the footage afterwards to perform the observer role.

They contracted Amey plc to build a visual AI model could perform that detection process more quickly, and in real time. Building that model required a large number of training images.

Using a few example images from dashboard cameras, we were able to generate the volume of data required, so the model could be trained and used.

SOMETHING ABOUT OUTCOME – time saved/efficiency etc