Visy has been delivering Rail OCR portals for DB Netze locations in Germany since 2018. One of the most important requirements set by the customer was to have extremely high quality images of passing cargo and automate data capture of wagons and cargo.

Today, Visy Rail OCR portals around Germany deliver to DB Netze high OCR hit rates of 98%, resulting in reliable and automated data processing workflows, as well as high quality images that are used in making repair assessments.

DB Netze AG is the rail infrastructure company of Deutsche Bahn AG in Germany. DB Netze is responsible for a rail network spanning 33,300 kilometers, making it Europe’s largest rail infrastructure provider. As a massive player in the rail space, DB Netze creates all the prerequisites for safe, environmentally friendly, and sustainable mobility in Germany and Europe.

As a new service, DB Netze wanted to offer its customers images of the passing wagons. One of their customers is DB Cargo, Europe’s leading freight railway. DB Cargo uses the images to digitize their maintenance processes especially to identify damage on their assets.

Before the implementation of Visy Train Gate systems, DB Cargo faced 3 main challenges:

1. Wagon numbers are often in bad condition or even obstructed and therefore difficult to verify for a human
2. The damage assessment process was slow, dangerous, and costly as it was performed by a clerk working on the tracks conducting physical inspections
3. A standard OCR solution could not meet the individual site requirements due to the limited possibilities of portal placement and the terrain which forced trains to vary their speeds

DB Netze selected Visy as its OCR and Train Gate provider in 2018, with the first site going live soon after in Nuremberg. After the system met DB Netze’s strict testing and performance criteria, Visy became its preferred OCR system provider, and the Train Gate solution was rolled out to six further locations: Hagen, Maschen, Mannheim, Munich, Seddin, and Seelze. More than 12 Visy Rail OCR portals are now deployed across these sites.

High recognition rates, even on damaged wagons

Visy Rail OCR portals deliver recognition rates above 98% at every location, built on Visy’s Deep Neural Network (DNN). For wagons with obstructed numbers, Visy added RFID, so that wagons carrying transponders can be verified automatically where a camera read alone could fall short.

Remote damage inspection

Each portal carries up to nine ultra-high-resolution cameras. These provide the foundation for OCR and let clerks carry out their damage inspections remotely, so cargo keeps moving and never has to stop for inspection. Reviewing images from a remote office, a clerk can confirm that a wagon is safe to travel or flag it for repair – inspecting far more assets per day than the old trackside process allowed.

Managing variable train speeds

Site conditions that affect train speed are handled by Visy Measurement feature. The laser-scanner-based system measures train speed in real time and adjusts the cameras and lighting to match. Train drivers operate as they normally would, with no need to account for the OCR system; the trains run safely and efficiently while Visy responds to their movement patterns and performs its imaging and OCR functions.

Direct data exchange with DB Netze’s systems

Interoperability with third-party systems is one of the most important features of the DB Netze installation. VisyXMLGate automatically shares wagon and cargo data with the InfraView system, exporting all Visy data directly into DB Netze’s own user interface. Staff continue working in their internal system rather than switching to third-party screens: their front-end stays under their control with the feature set they want, while behind the scenes it is updated automatically with data from the Visy system.

Local delivery and support

Visy’s German partner, LMT GmbH, was central to delivering the system and provides ongoing maintenance and aftermarket support. With a physical presence in Germany, LMT can respond to any DB Netze site within a strict SLA window. Because the engineers who deployed the system also support it, performance issues are resolved quickly – keeping system uptime high.