AI to identify wagons and containers: the new drive of the Port of Barcelona
Improve control of rail transport and make the Port of Barcelona a more sustainable, technological and smart environment. These are the main objectives of a pilot test of technology based on artificial intelligence (AI). It was carried on at the Hutchison Ports BEST for just over a year. The system of the startup AllRead Machine Learning Technologies (MLT) identifies in real time the containers and wagons entering the port. This improves the control of rail transport. This is another impulse for the Port of Barcelona to be at the forefront of the sector in the technological and digital field.
Reading precision technology
The startup AllRead, specialized in computer vision systems, has developed a system based on artificial intelligence. It allows to identify in real time the containers and wagons entering the port area. And this can be done without the need to install large infrastructures and hardware equipment. Simply, through a software capable of identifying numbers, letters and symbols in videos and photographs.
"It is a computer vision software applied to the detection and reading of texts, codes and symbols in supply chains," explains Adriaan Landman, Director of Operations and co-founder of AllRead. "Connected to any type of fixed or mobile camera, it improves the traceability of assets in ports, logistics platforms or transport companies, returning a record in real time and with great precision."
The technology was tested for approximately one year at the Port of Barcelona, where the pilot test analyzed a total of 950 trains and 13,500 containers. According to the AllRead co-founder, the neural networks were trained using images or videos taken by those responsible for the project, and by cameras previously installed in the port.
The results were very positive. Of the 13,500 containers that entered the Port and were analyzed in the pilot test, 99% were correctly identified. The technology was able to interpret license plates in vertical, horizontal and even combined positions.
It was also able to identify those that looked broken, dirty, and partially covered. The tests were carried out both day and night, to check the viability of the system in low light conditions.
Organization, cost reduction, and sustainability
For the Port of Barcelona, this identification system has numerous advantages. First, it improves the optimization processes. “By reducing repetitive manual tasks, such as manually managing the registration of each container and its position in the train car, we help increase the efficiency of operating teams. Better information management speeds up all processes without lowering quality,” explains Landman.
A direct consequence of this optimization is a cut in costs. AllRead technology reduces loading and unloading times and manual tasks associated with controlling the access to the terminal. In addition, its characteristics make installation and maintenance costs low. "It’s able to find the information in any position of the image, resisting dust, rotation, distance and damage, without losing reading quality. It does not depend on the quality of the camera, but on the robustness of the reading system," Landman points out.
This way, AllRead's software can work with the Port's existing equipment and be installed anywhere in its facilities. "By not requiring the additional implementation of hardware, cameras, firing lasers, lighting and other elements for the operation of AllRead software, the Port saves a significant investment in hardware, as well as maintenance costs," Landman adds.
Lastly, sustainability comes into play. Many states are promoting the use of the train over other more polluting means of transport. This meets the objective of reducing carbon dioxide (CO2) emissions and other greenhouse gases.
This is the road to achieve the goals set by the Paris Agreement. By streamlining operations, AllRead technology optimizes the use of the train in the port.
“Thanks to the Chief of Operations of the Barcelona Port Authority, we realized the high demand for these solutions in the port environment. In Spain, no port has applied an OCR [optical character recognition] that reads wagons and containers, while the percentage of cargo transported by rail grows every year," explains Landman.
Of the 13,500 containers that entered the Port of Barcelona analyzed in the pilot test, 99% of them were correctly identified
Barcelona is, today, a pioneer in the use of this technology in a port environment. Its implementation has made it possible to automate data collection, optimize operations and thus favor the loading and unloading processes. All this, with a fast and agile system and without the need to install large technological devices.
With its adoption, the port is positioned as a benchmark and leaves the door open for future improvements. On the one hand, and thanks to deep learning, the system can continue to improve its level of certainty. On the other hand, the analysis of the results allows it to continue innovating and developing solutions for the port environment.
"Now that we offer traceability of wagons, trucks, containers or dangerous goods, our interlocutors ask us to identify new elements, such as the identification of container seals or damages that could appear," says Landman. "As a technology company focused on the development of new computer vision solutions, we integrate these requirements into our roadmap."
The adoption of this computer vision and AI system is part of the Port of Barcelona's commitment to innovation and digitization. Thus, the institution joins a trend that is gaining weight worldwide.
Recently, the Port of Abu Dhabi announced a collaboration with Microsoft Corporation. The agreement seeks to improve container tracking and autonomous shipping capabilities at the Khalifa port. In this case, recognition will be made thanks to the artificial intelligence system of Azure (Microsoft's cloud computing service).
"The technology will drive the introduction of intelligent container tracking solutions based on artificial intelligence, which will guarantee 100% traceability of all containers handled, as well as the potential to launch an autonomous vehicle system," the Port of Abu Dhabi explains. They hope that its adoption will serve to modernize the port environment, reduce its carbon footprint, mitigate costs and optimize response times.
A pilot project has been carried out in the Port of Algeciras to test an innovative solution based on artificial vision and machine learning. The solution sought to automate the traceability and geolocation of ro-ro cargo within the Heavy Traffic Terminal. In this case, a network of video cameras deployed by the terminal captured images that were later processed and analyzed by the system.
"The ultimate purpose was to identify and monitor the trucks that pass through the TTP and automatically detect events of interest from the analysis of images and videos, such as, for example, the occupation or liberalization of parking spaces," the Port of Algeciras explains.
As it was the case of Barcelona, the Algeciras pilot project was considered a success. Everything indicates that, in the coming years, the identification of assets using AI and other technological solutions may become a fundamental service. These elements will improve the efficiency of port environments.