Shipping Container OCR

Add a reading layer to every gate, crane, and rail siding with shipping container OCR. Roboflow reads container numbers from the cameras you already have at gates, cranes, rail lines, and yard rows, validates every read against the ISO 6346 check digit, and turns every move into a verified record. One Roboflow customer runs container-yard OCR at 95 percent read accuracy across 26 million inferences a day, in production, on their own cameras.

Container Number Recognition:

• Read the full ISO 6346 container number, owner code, category, serial, and check digit, from doors, sides, and tops, in horizontal and vertical layouts, at gate speed and crane height
• Validate every read against the ISO 6346 check digit and your expected-container list, so a misread is caught by arithmetic before it reaches the TOS
• Read through the real world: faded and repainted numbers, rust streaks, dents, placards and stickers partially covering characters, night, rain, and headlight glare

Gates, Cranes, and Rail:

• Automate gate in and gate out: container number, ISO size-type code, chassis and license plate, door direction, and seal presence from one camera portal, without stopping the truck
• Read containers under the spreader at cranes and reach stackers, so the ID is confirmed at pick and set rather than called over the radio
• Read boxes on well cars and double stacks at rail speed as trains arrive and depart, so the consist is verified against the manifest without a walk

Yards, Inventory, and Systems Integration:

• Verify stack positions from yard cameras and camera-equipped equipment, so the TOS inventory reflects where boxes actually are, row by row
• Pair each read with damage, seal, and placard detection, so the gate record carries condition and dangerous goods context alongside the number
• Push verified reads into TOS, gate operating systems, and rail systems through API integration, with the image behind every read, every exception, and every dispute

Bring intelligence to every container move today.

‍

More About Shipping Container OCR

What is shipping container OCR?

Shipping container OCR uses cameras at gates, cranes, rail lines, and yard positions and deep-learning models to find and read container numbers, the eleven-character ISO 6346 identity with owner code, serial, and check digit, along with size-type codes, chassis IDs, license plates, and seals, and to validate every read before it enters the terminal operating system. Container number recognition runs at gate speed on moving trucks, at crane height under the spreader, and at rail speed on arriving trains, with the check digit and expected-container lists catching misreads by arithmetic, and every read logged with its image. It is the reading layer inside broader operations: for full yard visibility, slot occupancy, and dwell, see yard management and trailer detection; for barcode-based tracking inside facilities, see barcode scanning and tracking.

Does it need OCR portals and special cameras, or can it use ours?

It runs on the cameras you have, and that is the difference from traditional gate OCR portals. Purpose-built OCR lanes with line-scan cameras and controlled lighting work well and cost accordingly, per lane, per site; deep-learning OCR reads from standard IP cameras at gates, on cranes, along rail lines, and across yard rows, so coverage extends to every point a container is seen rather than the two lanes the portal budget bought. Existing gate and security cameras are often usable as-is, a camera added at a crane or a rail line is an ordinary industrial camera, and inference runs on an edge device on site at whatever volume the operation produces.

Can it integrate with our TOS and gate systems?

Yes. Roboflow Inference runs on edge devices at the site and exposes a standard API and webhooks, so verified reads flow into your existing systems: terminal operating systems like Navis N4, Tideworks, and Kalmar, gate operating systems and appointment platforms, rail management and consist systems, yard management and inventory, EDI flows and customer portals, and alerting into Slack, Microsoft Teams, and radio dispatch, through REST, MQTT, webhooks, and direct database writes.

‍