Text recognition (OCR) in plans

The goal of this AI service is to enable the capturing of metadata and semantic information preserved in plans.

In addition to the geometry of the building and its components, plans also contain semantic information that is recorded in text form (e.g., room information, dimensions). For the automatic analysis of scanned plans, these text elements must first be localized, then recognized and consequently processed.

The AI service “Text Localization in Plans” uses AI to localize text elements, which provides the basis for this service. Now this service can process the localized text elements and return them in machine-readable strings. High-performance optical character recognition (OCR) models are used for this purpose.

Data

  • Quality: All text snippet boxes should contain the text to be recognized, but as few surrounding pixel rows as possible.
  • Input data:
    • Image snippet with text elements (such as from BIMKIT service text localization).
      • JSON file
        • List of all recognized bounding boxes (per bounding box: coordinates for 2 points in 2D)
      • PNG files
        • Original file for further processing
        • Visualization of the result
  • Output data:
    • List with text elements

Contact Persons:
Phillip Schönfelder, M. Sc., Lehrstuhl für Informatik im Bauwesen, Ruhr-Universität Bochum
Björn Buck, A+S Consult GmbH (A+S)
Mike Nicke, A+S Consult GmbH (A+S)
Felix Kretschmann, elevait GmbH & Co. KG
Bianca Preißler, elevait GmbH & Co. KG