Your charger can't see who's hurting it.
A charge point knows its own electrical state and nothing else. It cannot tell that a petrol car has blocked the bay since morning, that a charged EV has been squatting for three hours, or that someone is walking away with its cable. RIOD Station Vision adds eyes to the site: edge AI cameras and AIoT sensors that watch the things OCPP will never report, and raise alarms while there is still time to act.
Four detection categories, all fused with charger state.
Cable cut and theft alarms
Copper charging cables are the most stolen component on Indian charging sites, and a cut cable means a dead charger, an insurance claim, and weeks of lost revenue. Station Vision pairs camera detection of tampering behaviour with AIoT sensing on the cable itself (continuity and motion). When both agree, the site raises a local siren, floodlight trigger, and an instant alert to your operations dashboard and WhatsApp, with the video clip attached.
Overstay and idle parking
A charged car occupying a bay earns nothing and blocks the next session. The camera watches occupancy; the charger reports session state. Vehicle present plus session complete equals idle. The system starts the overstay clock, notifies the driver through the app, and can trigger your idle fee automatically.
ICE-ing and wrong vehicle detection
Vision classification distinguishes an EV plugged into your charger from a petrol vehicle using the bay as free parking. Number plate recognition flags repeat offenders and non-customer vehicles. For fleet and residential sites, whitelist mode alerts on any vehicle that does not belong.
Mishandling and vandalism
Connector dropped on the ground instead of holstered. Cable driven over. Screen tampering. Physical impact against the unit. Camera events and the charger's own fault codes are correlated in one timeline, so when a connector fails in month eight you can see the three drops and one drive-over that caused it, and bill or train accordingly.
Edge AI, sensor fusion, and it works with what you have.
Edge AI, not cloud CCTV
Detection models run on low-cost camera modules at the site. No video streams to a server, no bandwidth dependency, and alerts keep working when the internet does not. Clips upload only when an event fires, which keeps data costs near zero.
AIoT sensor fusion
Cameras alone produce false alarms; sensors alone miss context. Cable continuity, motion, impact, and charger session state are fused with the vision events so an alarm means something actually happened. A stray dog near the pedestal is not an incident. A person handling a cable outside an active session at 2 am is.
Works with what you have
Alerts flow into the RIOD Operations Suite as tickets, into Charger360 dashboards, or to any third-party CSMS over API and webhooks. Sites with existing CCTV can often reuse mounted cameras where position and resolution allow.
Retrofit friendly
No civil work beyond camera mounting and a power point. A typical 4-charger site is live in a day.
Four site types, four reasons to deploy.
CPOs & public sites
Protect revenue per bay and stop cable theft from eating margins. Idle fees enforce themselves with photographic evidence attached, which ends billing disputes before they start.
Malls, hotels & paid parking
Your parking team cannot stand beside the chargers all day. Station Vision alerts the control room with a clip, so one guard covers every bay on the property.
Housing societies
The commonest society complaint is not electrical, it is behavioural: someone else's car in the designated slot, a resident hogging the shared charger overnight. Automated, evidence-backed alerts settle these without a WhatsApp group war.
Fleet depots
Wrong vehicle in the fast-charge bay delays the whole rotation. Whitelist enforcement and connector-handling reports keep depot discipline measurable.
Spec sheet
| Alerts | App, dashboard, WhatsApp, local siren/light trigger |
|---|---|
| Evidence | Event video clip + photo, timestamped, configurable retention |
| Detection set | Cable tamper/cut, overstay, ICE-ing, wrong vehicle, connector mishandling, impact |
| Integration | RIOD Operations Suite, Charger360, third-party CSMS via API/webhook |
| Connectivity | Works offline at the edge; syncs events when back online |
| Privacy | On-device processing, event-only recording, configurable plate masking |
Frequently asked, answered plainly.
Does this need a separate internet connection?
No. Detection runs on the camera module itself. Events queue locally during outages and sync when connectivity returns. A 4G option covers sites with no wired network.
Will it false-alarm on pedestrians and animals?
The models are trained to distinguish handling behaviour from passing traffic, and sensor fusion requires physical confirmation for high-severity alarms like cable cut. Sensitivity is tunable per site.
Can it enforce idle fees automatically?
Yes, when paired with the RIOD Operations Suite or a CSMS that supports idle billing. Station Vision supplies the verified overstay event; your billing system applies the tariff you configure.
What about privacy and CCTV rules?
Processing happens on the device and only event clips are stored. Plate data can be masked or hashed for residential deployments. Signage templates for camera notice are included in the deployment kit. Deployments follow applicable data-protection obligations for the site's jurisdiction.
Does it work with chargers that are not RIOD's?
Yes. The vision layer is charger-agnostic. Session-state fusion needs OCPP data from your CSMS, which most platforms expose.
Stop paying for what you can't see.
One stolen cable pays for a camera module several times over. Tell us about your site and we will spec the deployment.