OCR many parts

[PazottoPalma]

Hello!

I'm reading, using the IMAQ OCR read text vi, many parts of many images. For this, I make an array of diferent ROIs, each ROI for of information I want.

BUT, although the images be of the same pattern, some small diferences appears, and the array of ROI stays out of sync!

The question is: there is a another way to read many parts of a text ou image? As the OCR contained in scanners ou something like....?

Thank you very much!

Filipe Palma.

[BruceAmmons]

LabVIEW OCR is designed for reading a single line of text.  It doesn't do well at all on multiline documents.  It just isn't designed for it.  I would use the software that comes with a scanner if you want to use OCR on a large document.

Bruce

[PazottoPalma]

humm.. OK! thanks

[jc.benoit]

Hi,

I confirm the OCR has been designed to read one line at a time.

In order to detect each line, you may use the "Y Axis Averages" output of the "IMAQ Linear Averages.vi" (analysis palette).

Then for each line, build a rectangular ROI descriptor and use it as an input for "IMAQ OCR Read Text 3.vi".

Best Regards.

JC Benoit

Vision Engineer

Alliance Vision

www.alliancevision.com