Smart Port Turnaround Optimisation

Pacific Ports Authority was at capacity. Yard congestion meant vessels queued at anchor for 12–18 hours, costing shipping lines up to $80,000/day in demurrage. Containers needed an average of 2.4 re-handles to reach the correct vessel, and manual gate-in/gate-out data entry had an 8% error rate.

1. 1

   Installed 4K IP cameras on all 14 ship-to-shore cranes and 6 gate lanes; a YOLOv8 model reads ISO codes at 98.5% accuracy in all lighting conditions.

2. 2

   Built a Kafka event bus where each container scan generates an immutable event consumed by the yard-management optimiser and vessel ETA engine.

3. 3

   Developed a constraint-satisfaction optimiser for crane-move scheduling that minimises re-handles by sequencing moves against predicted vessel departure windows.

4. 4

   Replaced paper dockets with RFID-integrated truck booking slots, synchronising arrivals with crane availability windows.

5. 5

   Deployed a React + D3.js operations dashboard giving planners a live birds-eye view of yard occupancy, crane utilisation, and vessel time-to-departure.

We built an automated container-tracking and yard-intelligence platform. Computer vision reads every container ID as it moves — crane pick, set-down, gate transit — eliminating manual entry. This live position data feeds a scheduling optimiser that sequences all crane moves to minimise re-handles and maximise vessel throughput.