Why is there no output from the "IMAQ Co-occurrence Matrix" VI?

[Greg-H]

Hi,

It does not appear that the problem has been fixed in the latest release of IMAQ.  R&D is aware of the issue.  There is a post previously in this thread that provides a work around to this problem.

Regards,

Greg H.

[cplaw]

Unfortunately the work around is impractically slow, this problem is holding up part of the project that is migrating code away from Matlab. Is there an indication of when the fix will be available?

[GregSands]

I've just tried this with LV 2012 and Vision 2012, and do get a GLCM output.  I also used a displacement of 1,0 but set the co-occurence level to 256 (for an 8-bit image).  Maybe you can try again with the latest software.

There are also two other threads which give some pure LV code to produce a co-occurence matrix which appears to give identical results to the IMAQ function here:

http://forums.ni.com/t5/Machine-Vision/Texture-Analysis/m-p/193054

http://forums.ni.com/t5/Machine-Vision/Is-there-an-IMAQ-Vision-function-that-calculates-a-co-occurre...

There is also code there to compute the Haralick features, and it all seems fairly fast.

[cplaw]

Hi Greg,

That's good news, we have the "NI Academic Site License - College Teaching". Do you know when we can expect 2012 to be shipped to us?

Cheers

Colin

[GregSands]

No, I don't work for NI - just another user!

[cplaw]

Oops, my bad. Confusion between GregS and Greg-H 🙂

[Greg-H]

Hi,

I would suspect that you should have already received LabVIEW 2012.  If you have not you can contact your account representative to see why you have not received it yet.  You could also download it from our website.

Regards,

Greg H.

[cplaw]

I found who had been sitting on our copy...

Anyway the Co-occurrence Matrix is still zero length. Which makes me wonder if the upgrade upgraded this bit.

The vi I have is IMAQ Cooccurrence Matrix, Source version 11.0, Current revision 158.

Could someone please advise which version has the bugfix?

Thanks

Colin

[Ever_Song]

Hi Ryan,

Looking through the bug report (Corrective Action Request or CAR; number 308871) it looks like this bug was corrected in the 2012 (12.0) version of the Vision Development Module (VDM). You can see the 2012 VDM update log that lists this fix here.

[GregSands]

Vision 2012 is where I noticed the fix.