MobilityEnvironmentalLogistics

Automatic reading of seals

Artificial intelligence for identifying security seals on maritime containers

Parties involved

Promotors

Associates

Collaborators

Sustainable Development Goals

ODS

(Sustainable Development Goals)

ODS 1 ODS 1 ODS 1

Background

Maritime container traffic at the Port of Barcelona increased to more than 3.3 million TEU in 2019 and grew by 30.4% until July 2021 compared to the previous year. Container movement is organized through trains and trucks, with a tracking system that includes access control, partly automated by OCR technology, to verify security seals according to the ISO PAS 17712 standard. Although some ports have automated this process, most still require manual intervention to ensure container security, presenting a significant challenge in the management and control of seals.

Description

Barcelona Port Innovation, Hutchison BEST and AllRead have created a pilot project to monitor one of the entrances to the Port of Barcelona terminal. The aim is to automatically extract real-time readings from the security seals using a fixed camera installed at entrance 5 of the Hutchison BEST terminal.
Specifically, the project consists of:

  • Develop new Deep Neural Network technologies applied to the automatic control of security seals .
  • Design, develop and integrate a prototype of the system.
  • Carry out validation studies in the system’s real operational environment .

Conclusions

  • Evaluation of 60 traffic.
  • Although there is room for improvement, the technology works satisfactorily for this use case, as long as minimum image quality requirements are respected.
  • Many excluded traffics are due to the process inherent to this pilot. Cases such as containers that stop outside the image and rotated seals have an important representativeness and should be addressed differently in the event that this solution goes into production.
  • If you want to consider the different dimensions of containers , you should have at least 2 cameras/optics.
  • Even if software latency is greatly improved, ensuring that the camera generates quality images of moving seals is quite complex.

Project impact

Profit (Economics)

  • Improvement in the efficiency of operations : currently, it is the operators who do it manually, with a higher time and possible human error, which can cause delays in the entry and exit process at the terminals.
  • 90% improvement in efficiency in operating costs: savings of more than €1,250,000 per year with the use of software vs. manual process (Calculation bases in the report-annex Final report).

 

People

  • Improvement of worker health and safety : the manual control process is avoided, as this is slow and dangerous, generating a risk for operators who are between trucks and loads of more than 30 tons.

 

Planet (Environment)

  • 77,000 kg of CO2 eq. annually avoided from being emitted into the atmosphere: through the reduction of truck waiting times, idling at the terminal. BEST manages approx. 1M seals per year and where the manual validation process takes around 30 seconds.

News