Digitizing Seaports Using Optical Character Recognition (OCR) System Reduces Container Truck Waiting Time by 75%

BOTTOM BOTTOM AT SEAPORT CONTROL GATE The customer is a large-scale deep-water port operator, receiving more than 3,000 container trucks entering and exiting each day. The most painful problem of the port is the kilometer-long congestion at the gate control area (Gate-in / Gate-out).

According to the old process, when a container truck enters the gate, the control officer must run out of the duty station to visually check and record by hand in a hardcover book the information: Tractor license plate, Trailer license plate, and especially the international container code string (11-character string of letters and numbers). number printed vertically on the box). Then, the employee returns to the computer to retype this code into the port management system (TOS - Terminal Operating System) to authenticate the status of the goods. This entire manual process takes an average of 3 to 5 minutes per vehicle. During peak hours, errors due to typing in wrong codes occur continuously, causing goods to be sent to the wrong storage location, causing serious damage to the loading and unloading time of cargo ships.

COMPUTER VISION AND CHARACTER RECOGNITION (OCR) TECHNOLOGY A team of technology engineers has completely "smashed and rebuilt" the gate control process using the Optical Character Recognition solution (Optical Character Recognition - OCR) combined with Artificial Intelligence (AI):

  1. Installation of Multi-stream Cameras system: Creating a "digital scanning dome" at the port gate. Each lane is equipped with 6 high-resolution cameras combined with infrared lights. When the truck passes through the gate at a speed of less than 30km/h (no need to stop completely), the cameras will continuously capture images from the following angles: front (license plate), rear (trailer license plate), both sides and roof of the container (to capture the code and check the damage and distortion of the container).
  2. Training AI to read and understand Container codes (AI-Powered OCR): Different from static OCR software Previously (easily disabled by glare, dirt or rust on containers), the new OCR system uses a Recurrent Neural Network (RNN) model. The AI ​​is "trained" with millions of images of old, rusty containers with blurred or partially obscured text. The algorithm has the ability to extrapolate and accurately read the ISO standard 11-character code string even though the image is of extremely poor quality.
  3. Automatic comparison and barrier system: In less than 1.5 seconds, the code string read from the image will be sent directly to the port management system (TOS) by the API. The system will immediately compare: Does this truck have an appointment (Booking) at this time frame? If valid, the system automatically opens the barrier to let vehicles pass, and at the same time sends an SMS message (or displays on the electronic LED board) to inform the driver of the exact location of the block/Row/Tier they need to go to to drop off or tow goods.

OPTIMIZATION OF FLOW AND SUPPLY CHAIN SAFETY After only 4 months of implementing the Smart Gate OCR solution, management The port management has achieved unimaginable operational performance indicators (KPIs):

  • Processing time reached 0 seconds: Shortened gate passage time from 5 minutes to less than 10 seconds (vehicles move slowly through the camera dome), completely ending long traffic jams outside the national highway.
  • Data accuracy reached 99.8%: Completely eliminated the situation The situation of losing the location of a container in the yard due to an employee mistyping the code.

Digitization of identity proof: Images of the condition of the container shell (dented, scratched) are captured by AI and stored clearly at the time of entry. This helps the port eliminate 100% of legal complaints or unreasonable compensation requests from shipping lines regarding goods damage.