Vision Assistant OCR tool

[JerpVTX]

I am trying to read labels and some of the text is encroaching on neighboring characters.  For example, I have to "-V" but the "-" is withn the footprint of the "V" so OCR sees it as one character. 

How do I get around this?

Thanks for the help!

[JerpVTX]

Additional info; results from IMAQ Read Text 4:

Output string = 462-9103-V1-     Length = 12 (which I would expect)

Character Report only has 11 characters.  The 8th character is "-V"

In other words the VI Output string is not the same as the Character report output.

[BruceAmmons]

Try turning Auto Split on.  This will separate overlapping characters, according to the manual.

Bruce

[JerpVTX]

Thanks, I'll give it a shot.  Still don't understand why the 2 outputs from one VI aren't consistent though.

[JerpVTX]

Tried it but it didn't help.  Maybe I set it up wrong?!

[AnalogKid2DigitalMan]

Can you add the - as a trained character? If so and they are identified as unique entities can you just parse them out of your data?

-AK2DM

[JerpVTX]

Thanks for the suggestion, but the "-" is already trained AND being picked up twice (2 out of 3 occurrences) in the same inspection.  If you look at the length of the output string there are 12 characters, which is correct (9 alpha numeric and 3 “-“).  But if you look at the size of the character report, it is only 11.

[AnalogKid2DigitalMan]

Sorry, looking at the image closer I now see the red borders around the -.

Not sure about the count discrepancy, I have only dabbled in OCR with no deployed programs.

-AK2DM