The concept of Intelligent Video Gate (IVG) consists of a gate system installed at relevant railway nodes equipped with cameras and RFID readers for automatic identification of wagons and intermodal loading units (ILU; for example, containers, swap bodies, and semi-trailers) as well as damages, through image recognition and detection of wagon numbers, loading unit codes, placards and RFID units. The IVG is to be located at, or nearby, railway facilities where it can lead to significant improvements for processes within the supply chain, related to time, planning, work safety, maintenance and claims.

The aim of this deliverable is to describe a demonstrator of the IVG, a technique to enhance optimization of a fully operational terminal or yard, and with data management to enable fast and reliable detection of incoming and outgoing assets. Through automatic detection by an IVG of wagon numbers and intermodal loading units (ILU) handled, including recognition of dangerous goods, their sequence as well as visible damages, processes at terminals and yards can be optimized to achieve efficient dwell times and handling, as well as facilitate processes at other actors in the supply chain.

The R&I highlighted in this report are related to three tasks; Task 3.3: AI for images processing, Task 3.4: Data sharing and exploitation and Task 3.5: Demonstration and Evaluation. Task 3.3 and Task 3.4 have partially been reported in (D3.2, 2021) and further elaborated in this deliverable, Chapters 3 and 4. Task 3.5 entails the use cases considered in the exploitation plan, described in Chapter 5 and evaluated in Chapter 6.

Use cases evaluated in this deliverable are dependent on correct extraction of information from the images produced by the IVG, as well as on reliable storage and sharing of the resulting data. The current accuracy levels for the IVG of Trafikverket in Gothenburg, Sweden have been demonstrated in this deliverable. The evaluations that have been performed show that there is room for improvements of the information extraction, but also that the concept works well and that the extracted information has been successfully made available and makes several important use cases possible. The uses cases have also been demonstrated in the final event of FR8RAIL III on 2022/12/09.

It can be concluded that the IVG concept shows potential of automation and digitalization regarding reduction of time and speed-up of the technical checks on departure and arrival, use of image processing combined with machine learning as well as sharing and exploitation of data along the supply chain. However there are potentials for further improvements of the image processing and data sharing, to achieve accuracy levels above 95% through e.g. ability to recognize more code types (mainly national), using colour cameras and exploring common format for data sharing and semantics.