Road maintenance is an essential process for guaranteeing the quality of transportation in any city. A crucial step towards effective road maintenance is the ability to update the inventory of the road network. We present a proof of concept of a protocol for maintaining said inventory based on the use of unmanned aerial vehicles to quickly collect images which are processed by a computer vision program that automatically identifies potholes and their severity. Our protocol aims to provide information to local governments to prioritise the road network maintenance budget, and to be able to detect early stages of road deterioration so as to minimise maintenance expenditure.