Geye / Industry playbookPlatform pressure
A service gap changes how long passengers remain on a platform.
Geye / Industry playbook
Public Transportation
Review platform, corridor, and entry activity to understand flow and operational exceptions.
Planning guidance, not a claim of installed capabilities or customer results. Validate scope and feasibility before deployment.
Industry overview
Stations experience short demand surges tied to service arrivals and disruptions. Platform and concourse observations can help operations teams understand flow while keeping dispatch and safety decisions with trained personnel.
Concourse flow
Arriving and departing passengers can cross in opposite directions.
Accessible routes
Luggage or temporary barriers may narrow a designated access path.
Use cases
01
Platform pressure
- Context
- A service gap changes how long passengers remain on a platform.
- Human action
- Review platform occupancy during service gaps.
- Intended outcome
- A controller review informed by service conditions, not dispatch automation.
02
Concourse flow
- Context
- Arriving and departing passengers can cross in opposite directions.
- Human action
- Compare gate flow with concourse congestion.
- Intended outcome
- A directional-flow comparison for station operations.
03
Accessible routes
- Context
- Luggage or temporary barriers may narrow a designated access path.
- Human action
- Observe persistent obstruction near accessible routes.
- Intended outcome
- A route inspection request with location and observation time.
AI vision capabilities
Read the scene, retain the context
Evaluate crowd bands, bidirectional line counts, and dwell near marked access routes. Train occlusion, platform curvature, and mixed-direction movement require separate validation; no passenger identity tracking is needed.
Evidence contract
Keep zone, source timestamp, and view availability together. A stale frame must not become a current event.
A person owns the decision
A controller review informed by service conditions, not dispatch automation.
Define acceptanceProposed capabilities must be tested against representative inputs. Unsupported or low-confidence results belong in a review queue, not an automatic decision.
Real-world scenario
Hypothetical, not a client story
Hypothetical: a delayed service increases platform occupancy. A station controller checks adjacent cameras and field reports before applying the existing crowd-management procedure.
- 01
Context
A service gap changes how long passengers remain on a platform.
- 02
Review trigger
A candidate observation in the selected zone requires confirmation against current conditions.
- 03
Human response
Review platform occupancy during service gaps.
- 04
Intended result
A controller review informed by service conditions, not dispatch automation.
How it works
CCTV
Read an authorized stream from the selected zone. Track source health and timestamps; a missing view is not an empty scene.
AI analysis
Apply validated models and zone rules on assessed compute. Group candidate events and expose uncertainty instead of implying certainty.
Reviewed insight
An assigned person checks evidence, records the outcome, and follows the existing operational procedure. Rejected signals feed back into evaluation.
Business and operational value
Compare counts and dwell estimates with sampled observations across weather and service conditions.
Evidence quality
Accepted observations / reviewed observations. Count missed cases separately against the manual reference.
Target: agree before pilotReview effort
Record minutes per reviewed item, including rework and escalations. Compare the same task with the manual baseline.
Target: agree before pilotSafe failure
Record whether stale inputs, denied access, and unavailable sources stop or visibly degrade the workflow.
Target: agree before pilotEstablish a manual baseline, agree acceptance thresholds with the operational owner, and compare review effort as well as accuracy. Any benefit must be measured in the pilot; no savings or ROI are promised here.
Deployment plan
Pilot one platform and its entrance with station operations. Sample peak, off-peak, disruption, and wet-weather conditions; rehearse degraded coverage without changing train-control systems.
- Scope: name the owner, permitted inputs, reviewers, and acceptance criteria.
- Prepare: assess edge, on-premises, or approved hybrid compute, storage, power, and network capacity with representative streams.
- Validate: compare with human-labelled samples, exercise denied access and outages, and document rejected results.
- Decide: approve a limited rollout only after review; keep a rollback owner and re-evaluate when inputs change.
Existing CCTV and infrastructure
Check platform sightlines, vibration, lighting transitions, clock synchronization, and station network capacity. Verify authorized recorder streams separately from signaling infrastructure.
| Check | Required evidence |
|---|---|
| Source access | Authorized stream, codec, resolution, frame rate, recorder connection limit. |
| Scene quality | Day, night, occlusion, and missing-frame samples from the selected zone. |
| Compute & recovery | Concurrent decoding benchmark, storage budget, clock sync, outage fallback. |
Existing CCTV may be reusable after compatibility assessment, not by default. Inventory codecs, resolution, frame rate, authorized stream protocols, recorder limits, time synchronization, and sample quality. Benchmark concurrent decoding and models before sizing hardware.
Integration planning
01
Approved source
02
Scoped processing
03
Human approval
04
Controlled export
Example: compare reviewed occupancy intervals with an approved timetable feed. A control-room event queue needs agreed station identifiers and stale-data handling; no live dispatch integration is claimed.
These are example integration plans, not live connectors. Confirm schemas, least-privilege credentials, delivery acknowledgements, retry limits, and duplicate handling in a sandbox before enabling data exchange.
Security and privacy
Avoid linking movement to fare-card identities or tracking journeys across stations. Keep access-route observations aggregate and do not infer disability from appearance.
Least privilege
Approve reviewer roles and test a denied-access case before launch.
Data lifecycle
Set retention, deletion ownership, and encryption for stored and transmitted evidence.
Accountable operations
Log access and decisions; rehearse incident escalation and rollback.
Before launch, approve purpose and lawful access, role-based permissions, encryption configuration, retention and deletion rules, audit logging, and incident ownership. Verify these controls in the chosen environment; this page does not claim compliance certification.
Hypothetical pilot example
HYPOTHETICAL PILOT
One bounded question. Evidence before expansion.
Hypothetical pilot: staff annotate crowd bands over several service arrivals and one simulated feed outage. Accept expansion only if stale views are conspicuous and alerts support rather than overload the controller.
A proposed evaluation exercise, not a deployed case study. There are no named clients, claimed results, or implied endorsements.
Frequently asked questions
Does it control train movements?
No. Signaling, dispatch, and safety procedures remain separate.
Can it track a passenger journey?
That is outside this scope; observations are aggregated by approved station zone.
Will rain affect counts?
It can. Validate wet-weather scenes, umbrellas, glare, and low visibility before operational use.
Plan your next step
Bring an authorized camera inventory, representative scenes, and one operational question. Prepare a planning brief, then validate sizing assumptions with a scoped pilot.