Optical Character Recognition

Startek Technology can provide optical character recognition (OCR) software or associated OCR engine libraries. The OCR was originally developed for reading postcodes in the freight handling industry. The OCR software uses various digital processing methods to locate a connote postcode within an image, determine its orientation, and finally extract the postcode numbers for automatic sorting.

Postcode Reading Application

In this particular application large images of parcels (up to 6000×6000 pixels) were presented to our system via a CameraLink interface with timing data via a serial communications protocol. The consignment note postcodes had both machine printed addresses with post codes on a white label, as well as hand written postcodes in a boxed area. The sortation system has a secondary "Visual Coding" fall-back for when the postcode could not be attained automatically — in this case the parcel images are sent to the computer screen of one of several operators for them to read the postcode in the image and enter it into the system so that the parcel can still be sorted correctly by the sortation conveyo

OCR Demo

Please note the following image restrictions are enforced for this demo:

  • Well spaced, machine printed numbers only (No letters supported in demo)
  • Black writing on white background (dark on light, greyscale acceptable)
  • Must not contain noise or garbage (no lines, no frames, no non-numeric characters)
  • Must be a Jpeg or Bitmap (BMP) file
  • No wider than 600 pixels
  • No higher than 400 pixels
  • No more than 250kB file size
OCR Sample Image
OCR Sample

Upload file here to perform the OCR