Article • 03.06.2024
AI-supported detection of hazardous goods
Waterway police enhance security in the Port of Hamburg
Not all dangerous goods handled via the Port of Hamburg are declared as such and thus pose a safety risk for both transporters and the port as a whole. The waterway police have always been responsible for identifying undeclared dangerous goods and ensuring that they come into compliance with regulations. Previously, these inspections were carried out manually and, owing to the vast number of containers, only on a random basis.
The AI project was funded by the InnoTecHH Fund of the Hamburg Senate Chancellery.
Client: Hamburg Police
User team: Hamburg Waterway Police
Developer team: DAKOSY (GEGIS)
Funding body: Senate Chancellery of the Free and Hanseatic City of Hamburg (InnoTecHH Fund)
Consulting partner: dida Datenschmiede from Berlin
Lutz Dreyer, head of division WSP 52, the waterway police's Central Dangerous Goods Monitoring Unit, explains: "This has been an issue for us for decades. The sheer volume of containers handled via the port made the task as difficult as looking for a needle in a haystack. We also lacked the legal authority to inspect containers that were not declared as dangerous goods. We only received this legal authority following an incident in the Port of Hamburg in late summer 2016, when a container ship burned for three days because 33 containers of paint and thinners were not declared as dangerous goods."
In the wake of this event, the revised Hamburg Port Security Act (HmbHafenSG) of 11.05.2021 was enacted, including Section 25 (data processing to identify undeclared dangerous goods) under which the competent authorities (here: the waterway police) are authorized to process cargo data in order to determine whether undeclared dangerous goods are being loaded or unloaded to transport units in the Port of Hamburg. As the operator of the Port Community System for the Port of Hamburg, DAKOSY has access to all import and export shipment information about the containers which are handled. A technical analysis in 2021 showed that the search for undeclared dangerous goods could be considerably improved and made more efficient by incorporating intelligent processes, particularly from the field of artificial intelligence (AI). "With the financial support of the InnoTecHH Fund* of the Senate Chancellery of the Free and Hanseatic City of Hamburg and the technical support provided by DAKOSY, we started the AI project at the beginning of 2023," says Dreyer and continues: "During the first part of the AI project which ran through the end of 2023, we only worked with data from the IMP Import Management Platform which DAKOSY had integrated into our GEGIS system. The AI training was carried out in such a way that shipments were marked with a probability factor for undeclared dangerous goods so that we could then check them."
In the second part of the AI project, which was completed at the end of April, the functionality was extended to export data from the Export Management Platform. Thilo Buchheister, who manages the project for DAKOSY, describes the procedure: "To optimize accuracy, we moved from a random forest model to a convolutional neural network. In order to ensure that additional new product descriptions are also incorporated into the analysis, we have implemented an automated regular training process."
For Dreyer, the project objectives have been fulfilled: "We have successfully completed the project with DAKOSY and our GEGIS now has a continuously adaptive AI system. As a result, we receive effective recommendation lists that our staff can use to detect undeclared dangerous goods quickly and efficiently. With the help of automated training runs, our database is continuously expanded and kept up to date. And, of course, we remain in contact with DAKOSY regarding potential technological innovations."
*Thanks to the InnoTecHH Fund, the Senate Chancellery supports Hamburg authorities in testing and implementing innovative ideas (with a focus on artificial intelligence). The aim is to create tangible added value for the employees of Hamburg administrative agencies and for citizens in an agile and accessible way. The Department for IT and Digitalisation (ITD) has been providing funding and expertise for this since the start of 2023.
"We have successfully completed the project with DAKOSY and our GEGIS now has a continuously adaptive AI system."
