Netzgesellschaft Düsseldorf mbH

Troubleshooting and Diagnostics Assistant App (FDA)

The Troubleshooting and Diagnostics Assistant App (FDA) is based on our Medialesson proof of concept from 2021/22 and features a consistently modular architecture.

Thanks to clearly defined interfaces, new readout functions can be added with minimal development effort, while the white-label capability enables NGD customers to fully adapt the app to their own brand identity. This structure makes the application not only scalable and easy to maintain, but also extendable in the long term - providing ideal conditions for secure and sustainable use across a wide range of diagnostic scenarios.

Netz Düsseldorf
Netz Düsseldorf - FDA
Highlights
  • Automated extraction of measurement values from dot-matrix displays using customized OCR and advanced image preprocessing techniques
  • Robust session and state management allows recorded measurement values to be stored locally and transmitted later, ensuring reliability even in environments without internet connectivity
  • Machine packages are downloaded and managed using Unity AssetBundles, enabling the seamless addition of new machines without the need for app updates.
Netz Düsseldorf - FDA
Realization and Development

The FDA app uses AR image targets to automatically recognize the meter box and overlay an interactive guide directly on the relevant control elements. This enables the readout process to be carried out across a wide variety of devices from different manufacturers, without requiring users to learn how to operate each machine individually.

Since most of the environments offer only weak or no internet connectivity, it was essential to perform text recognition locally on the device—rather than using the Azure Computer Vision Toolbox as initially planned. To achieve this, extensive evaluations were conducted to establish optimal starting conditions. Specific parameters were identified that can be adjusted based on the machine type, ensuring reliable text recognition even in highly diverse environments.

Services
  • Design and architecture of the platform as a proof of concept

  • Further development of the proof of concept into a unified, extensible platform and white-label-capable app

  • Migration of the text recognition engine for displays from Azure Computer Vision to local Tesseract OCR via EmguCV

  • Integration with a custom backend to simplify the management of app updates for future machines

Technologies
  • Tesesseract OCR via EmguCV

  • ZXing

  • Unity AR Foundation

  • Unity UI Toolkit

  • Azure App Service

  • Azure Blob Service

  • Azure App Insights

  • .NET / C#

We'd love the opportunity to work together.