VLM output
At a sunny waterfront wharf, a single person in a white shirt crouches at the edge of a paved quay beside a black expandable barrier fence, then lies face-down on the asphalt next to the fence before starting to rise again. No vehicles, vessels, or other people are active in the immediate quay area; the harbor with cranes, a red cargo ship, and moored yachts is visible in the background.
- Person In Restricted Quay AreaYes
- Person Down / FallenYes
- Person Count1
- Barrier Fence IntactYes
- Barrier CrossedNo
- Cargo Handling ActivityNone
- Vehicle Present In ZoneNo
- Vessel Berthed At QuayNo
- VisibilityClear
Customer challenges
Large port areas and multiple agencies leave monitoring scattered across separate systems with no shared picture
Safety incidents rely on manual patrols, surface only after the fact, and accountability is hard to establish
Terminal peak crowds, boarding flows and shuttle dispatch are judged on site without real-time data
Legacy video AI needs one model per scenario, so every new requirement means new development
Berth utilisation and berthing efficiency lack objective measurement; scheduling runs on experience
Incident records are fragmented and cannot form auditable governance data
Customer benefits
24x7 AI-driven proactive port safety
Safety incidents shift from post-hoc review to second-level alerts
Real-time data for terminal crowds, flows and shuttle dispatch
Dozens of scenarios on one platform, cutting rollout and expansion cost
Berth and berthing efficiency measured, moving scheduling from experience to data
Rapid deployment on existing cameras, keeping adoption cost low
Key features
Port Operations War Room
Brings CCTV, AIS, IoT and port-centre data onto a single timeline, with a cross-agency dashboard, KPI monitoring and automatic recording of incident accountability.
Prompt-Defined VLM Detection
Raises man-overboard alerts in 3 seconds and detects restricted-zone intrusion, helmet and vest compliance, oil spills, abandoned objects, and passenger-terminal crowd density and abnormal flow, with new scenarios added by prompt and no retraining.
Operations Analytics
Logs berth occupancy and berthing operation times automatically, analyses berthing efficiency, supports berth scheduling and resource dispatch recommendations, and forwards data 24x7 to MQTT, API or database.
Port Digital Twin War Room
Runs the war room view on the port digital twin.
Before & after
| Before | After — with Linker AI Nexa |
|---|---|
| Before Each agency watches its own monitoring screens | After — with Linker AI Nexa A single port digital-twin war room |
| Before Safety incidents found by manual patrol, after the fact | After — with Linker AI Nexa Man-overboard, PPE and intrusion alerted within seconds |
| Before Terminal crowds judged by staff on site | After — with Linker AI Nexa Real-time crowd density and abnormal-flow detection |
| Before One model per scenario; new needs mean new development | After — with Linker AI Nexa Scenarios added by prompt, many tasks on one hardware stack |
| Before No measurement of berth use or berthing efficiency | After — with Linker AI Nexa Automatic logging with efficiency analysis driving scheduling |