Computer Vision Site Monitoring
Technical
THE CHALLENGE
A UK charge point operator running a network of charging hubs relied on manual checks for every operational incident, from session and payment faults to vandalism, accident damage, littering and site defects. Each check meant reviewing the CPMS and requesting CCTV clips, often followed by staff or third-party visits that added unforecasted operational spend. OCPP and OCPI data showed only part of the picture, and asset management reports quickly went out of date as site conditions changed.
OUR APPROACH
Deployed computer vision sensors across the charging hubs with a number of suppliers, overlaying operationally relevant parameters onto the live camera feeds
Implemented automated alerting on those parameters, with email alerts to asset managers so issues were flagged without manual monitoring of the network
Used custom sensors and edge computing to iterate alongside the operations team, building up a remote picture of each site issue before deciding on a visit
Enabled the operations team to dial into live feeds alongside engineers or site visitors and capture recorded footage as evidence for insurers or emergency services
Matched and verified camera data against OCPP and OCPI feeds, identifying vehicle types, customer groups, usage patterns and recurring fraudulent activity across sites
OUTCOME
Site visits dropped from quarterly to annually, cutting operations travel and visit costs by up to 75%, and time to fix non-charger issues fell from weeks to days. Forecasted operations headcount was reduced by 25% through autonomous monitoring.