Optical Character Recognition (OCR)

OCR stands for Optical Character Recognition, also known as text recognition. An OCR software can recognise printed texts on digital images and convert those texts into a text document. When the text from a document is typed manually, this “conversion” takes place through the person typing in front of the monitor.

OCR software is available as PC software, as a free online tool or as an or app for smartphones and tablets. However, printers, scanners and other multifunctional devices with integrated OCR software are now also available on the market. The selection of the correct solution depends on the scanning volume and the desired scope of functionalities, e.g. whether handwritten texts are to be recognised as well.

Advantages of OCR

  • Faster work process through automated data processing
  • Increased productivity of the work processes
  • Less potential for error than with manual entry
  • Automated allocation of scanned documents, images, etc.
  • Easier searching of images as well

Basis for OCR

The basis for Optical Character Recognition to function properly is, above all, the quality of the original/document. That includes, for example, the colour, contrast, layout and font. Furthermore, the document should not have any creases, stains, marks, strikethroughs or underlining. In the case of a manual photograph of the document, ensure that the lighting is good and you are holding the camera steady and straight above the paper.

How does OCR work?

  1. Document capturing: The page is captured by photograph or scanning and saved as an image file (raster graphic).
  2. Layout analysis of the software: Individual parts such as images, graphics or text blocks and their placement are recognised. All text blocks are automatically divided into paragraphs, sentences, words, and characters.
  3. Character recognition: In this step, methods of pattern and property recognition are used. The individual recognised pixels are compared with a database of patterns for letters/characters and computer fonts. An algorithm decides which pattern matches the character best. Then the software reinserts the recognised letter, number or punctuation mark in a common text encoding.
  4. Composition of the text: In the following step, the individual characters and letters are combined to form words and sentences again. Ideally, the software uses an integrated dictionary and spelling correction during this step so that logical sentences are created in the end. That way, entire paragraphs and the entire text are constructed bit by bit.
  5. Document saving: In the end, the text-recognition software saves the recognised texts in an editable file.

Areas of application/examples of OCR

  • Capturing addresses for postal consignments
  • Further processing of image files or making it possible to electronically search image files
  • Recognition of properties for electronically sorting documents
  • Identification of vehicle licence plates in traffic monitoring
  • Extended full-text search for contents (also in PDFs and images)
  • Support for the blind with voice output
  • Scanning in apps (e.g. invoice scanning in banking apps)
  • Translation app via photo/camera functionality

In the transport management solution CarLo as well, OCR is implemented for the logistics and transport documents. For example, the function allows you to directly read out transport orders from PDF files and enter orders fully automatically. Thus, manual entry is no longer necessary, which saves logistics and transport companies a great amount of time. The OCR component developed by Soloplan supports PDF files and even reads out texts from e-mail files.