Bus Shelters and Transit Stops, None of Them Staffed

Thousands of sites, and nobody standing at any of them. A bus network is not one estate with many cameras: it is many estates with a few cameras each, no staff on site, and a connection that may or may not be there, which is a deployment problem before it is an analytics problem.

The Constraint

Connectivity Decides the Design

A metro station has hostile lighting and excellent fiber. A bus shelter is the opposite: daylight is usually fine, and there is no fiber at all. That single difference decides the architecture, which is why this estate gets its own page rather than a paragraph on the railways one.

The answer is a single cost-effective edge server at the stop. Analytics run locally and only events cross the link, so bandwidth stops being the limit. When the link drops the site keeps recording and detecting, and synchronizes automatically once it returns — the failure mode of a dispersed estate is a site going quiet without anyone noticing, and that is the failure this design removes.

The detectors themselves are the ones the rest of the network already runs. Nothing on this page is a bus-specific product; it is the same catalog, configured for a site nobody is standing in.

Coverage

What an Unstaffed Stop Needs to See

Vandalism and Graffiti

Damage is expensive because it is found the next morning. A detector written as a sentence fires on the act instead. Beta, selected instances.

Explore StreamVLM™

Loitering and Dwell

Dwelling where and when dwelling is itself the signal — an empty shelter at 3 a.m. is a different rule from a queue at 8 a.m.

See Perimeter Intrusion Detection

Unattended Items

A bag or a box left at a shelter, on the same detector the interchanges run and set for a site with far less passing through it.

See Abandoned Object Detection

Camera Integrity

On an estate nobody visits, an obstructed or repositioned camera is an unmonitored site. A sprayed lens or a knocked mounting raises its own alert.

See Camera Tamper Detection

Weapons and Assault Precursors

Weapon detection at shelters and interchanges, routed to the operator with the frame attached rather than discovered afterwards.

See Weapon & Gunshot Detection

Bus Lane and Stop Obstruction

Vehicles stopped in a bus lane or across a stop box, and motorbikes using the lane, detected as a traffic-enforcement violation with an evidence package.

See Multi-Camera Traffic Analytics

Facts

Per-Site Hardware
A single cost-effective edge server covers a small stop; high-availability edge is available where a site cannot afford to go dark.
Bandwidth
Only events cross the link, not raw video, which is what makes a cellular or shared connection viable at all.
Offline Resilience
The site keeps recording and detecting when the link drops, and synchronizes events automatically on recovery.

FAQ

Our stops have no fiber. Is that a blocker?

No, it is the normal case for this estate. An edge server runs the analytics at the stop and sends only events, so the link carries a fraction of the data raw video would need. It keeps recording offline when the link drops and synchronizes automatically on recovery.

Do we need new cameras at every shelter?

No. Any camera that supports ONVIF or streams RTSP with H.264 or H.265 on a static IP can be connected, and each module then sets its own resolution and placement requirements, which a site survey confirms. Connection is not plug-and-play: a qualified network engineer configures each camera and router.

How is this different from the metro deployment?

The detectors are the same and the architecture is not. A metro network is a small number of dense, well-connected sites; a bus network is a large number of sparse, poorly connected ones. The railways page describes the first, this page the second.

Who owns the data?

You do. Customers retain 100% ownership and control of their data, and IREX neither owns nor accesses customer video, events, logs, watchlists, or floor plans. On-premises deployment is the data-residency mechanism.

How do we start?

With a pilot: one site, three to five use cases, six to twelve weeks, and two or three measurable success criteria agreed in writing before it begins. Accuracy is benchmarked on your own cameras, and the measured numbers go into the contract.

Start with One Corridor

One bus corridor, instrumented end to end, answers the connectivity question for the whole network far faster than a paper study.