Problem with advanced threshold in OCR - Vision Assistant 2013

[Ahmed_Osman]

I'm facing a problem with Vision Assistant 2013

The OCR charter set file advanced threshold data is always fixed at 255 upper value and optimize for speed is checked.
I edit them and reopen the file but no change.

anyone facing the same problem ?

[Spicy_Mystery]

Hi Ahmed,

I've replicated what you're experiencing and I think it's strange as well - it should be saving those settings.  I notice that under the regular Threshold tab, if I set it to manual and change the values and save, it will set it back to Auto, although it will keep the values.  But I can't save the Advanced Threshold values at all - they go back to the default, just as you said (for both min and max, and even the Optimize checkbox as well).  I am going to look into this further and speak with a few people to see if this is intended or if it is a bug.  Thanks for your post.

[Spicy_Mystery]

Hi Ahmed,

I've gotten a chance to look into this issue and it is definitely a bug in Vision Assistant 2013, and has been solved in the recently released Vision Development Module 2013 SP1.  You can download SP1 here:

http://www.ni.com/download/vision-development-module-2013-sp1/4566/en/

[Ahmed_Osman]

Thank you pro 🙂

[Ahmed_Osman]

This update fixed saving the ocr advanced threshold data.

but there is another problem the advanced threshold options is not applied on the main window, it's been applied inside edit ocr file only.

[Spicy_Mystery]

Hi Ahmed,

I'm not seeing this behavior, although I think I know why you might.  Setting them to manual is required, and it seems like it resets sometimes.  But I have it working alright.  Also, in the edit OCR file screen, make sure to hit enter so it enacts your changes, then save.

[Arroyo]

Dear all,

I am having a difficulty running the OCR in labview. If I run the algorithm in the vision assistant I obtain certain results (expected ones). 

If I generate a VI from the VIsion Assistant with exactly the same parameters and run it in LabView, I get other results (usually it does not identify the characters or find other ones). 

Does anybody know which could be the reason of this behaviour?

Thanks, 

Esteban

[FuManWho]

Hi Esteban,

One thing to take a look at is the example VI for OCR. This can be found by going to Help>>Find Examples in LabVIEW. From there navigate to Toolkits and Modules>>Vision>>OCR>>OCR.vi. This VI will give you an overview of how to use the OCR functions and how a sample program would look like. There may be things that are configured in your vision assistant that are not the same in the program. But if you would like further assistance on this issue, I would recommend starting a new thread as you are replying to a thread that is a year old. Newer threads will get more views.

[Arroyo]

Hi Paolo,

Thanks for your answer. Yeah I have seen the examples and I´m familiar with the use of the OCR VI. I have use it a couple of time already with good results.

My problem came last week. While I was trying to run OCR on my image in my LabVIEW code. The algorithm did not detected what I was expecting.

I decided to run the test on the vision assistant and it worked perfectly. I assumed my code had different parameters, so I decided to generate a VI from the vision asssistant and run it on LabVIEW for the same image to verify.

I did not change anything on the VI (all parameters are the same), and used the same image. Surprinsingly the results are different between the assistant and the VI. That strikes me a lot!

I´ll start a new thread as you recommended. Hope to find the solution soon. Thanks  

Regards,

Esteban