Technical Notes, Zetadocs

HOWTO: Ensure barcode values are detected correctly by OCR in the Zetadocs Server Document Converter

ZTN4091

ID: ZTN4091

This Zetadocs technical note applies to:

  • Zetadocs Server - Document Converter service 1.1.31 or later

Summary

Documents converted with OCR barcode detection enabled sometimes do not have their barcodes detected or produce incorrect values for detected barcodes. Barcode detection is highly dependent on the image quality of the barcode and the barcode font that is used. Incorrect barcode value detection can be caused by poor or scratchy scanned images, scanning at too low a resolution, barcode fonts that are too small, barcodes positioned on the page at an angle rather than horizontal or barcodes that contain unusual characters.

Recommendations

To increase the percentage of correct barcode values that are recognized ensure you do the following:

  • Included a star (*) character at the beginning and end of the barcode value.
  • Only use characters which are supported by your chosen font. See barcode specifications for details. We recommend that where possible you stick to simple uppercase alpha numeric values (A-Z, 0-9).
    • In some fonts lowercase characters are not supported and may be represented by a plus (+) sign and the uppercase equivalent letter.
  • Use a reasonable barcode font size. We recommend 32pt for our suggested font. See Zetadocs for NAV SDK Guide for more information.
  • When scanning use a resolution of at least 300 dpi.
  • Place barcode horizontal on the page rather than at an angle and in a location where it is unlikely to get written over.
  • Ensure there is at least 0.75 inch of white space around the barcode.
  • Ensure the scanner device does not save highly compressed image files (eg. JPEG) within the generated PDF, instead configure the device so that it saves in TIFF format instead (with loss-less or lower loss compression).
  • Some applications such as Microsoft Word include hidden characters such as carriage returns which can get included in barcode values and prevent it being detected. It is recommended that, where possible, documents should be printed to Zetadocs PDF and processed from there as these characters are removed.

If, despite following all the instructions above, you are still experiencing problems with a barcode, where possible, print it to Zetadocs PDF and process it from there. This seems to resolve the majority of issues with problematic barcodes.

References

Barcode Font Types

Last updated: 24 Sep 2015 (DC)