systematic mask bugs in Vision

[Root_Canal]

I keep running into a recurring bug in Vision functions. If you use an image as its own mask, some Vision functions fail to work properly (they ignore the mask).

Here is an example that shows the problem in three vision functions- IMAQ Histogram, IMAQ Convolute, and IMAQ Quantify.

 

 

-root

Message Edited by Root Canal on 05-01-2009 05:25 PM

[Root_Canal]

Here's the test image.

[SMeserve]

Hi Root,

It seems that this issue is an exact duplicate of the issue raised in this thread.  A corrective action request has been filed on this (#162298) and you can check in the next version of Vision to see if the CAR has been resolved by checking the release documentation for information based on the CAR ID number I just gave.  However, for now, the appropriate workaround is to create a copy of the image, as you have done, and use that as the mask.

[Root_Canal]

Shortly after mistakenly posting the question to the hardware forum, I sent a message to the moderator and asked that the post be removed as it was a duplicate post. I then posted the question to the Labview Software forums, as it is a software issue, not hardware. I do not know why the moderator has chosen not to delete the duplicate post.
 
The issue has not been resolved. A support engineer merely promised to document the bug in one VI. He did not acknowledge that this bug is present in more than one VI, or that it was a bug instead of a documentation issue.

I assumed that the support engineer misread or misunderstood my post beacuse it is obvious that no programmer would want a VI to behave in this fashion. If someone wants the VI to analyze the whole image, they simply leave the mask input unwired, just like the Vision documentation tells them to.

Multiple VIs have this issue and it is quite evidently a bug. Please delete any duplicate posts and post an official response regarding the presence of the mask bugs.

Thank you,

-root

[Intaris]

The CAR mentioned int he previous post is the internal bug-tracking method used by NI.  If you have a CAR, it's filed as a bug.

Shane.

[Root_Canal]

Shane, this is a quote from the support engineer who created the CAR:

"

Root Canel [sic],

If you wire the same image into both the source and the mask of and IMAQ Analysis vis the vi analyzes the entire image. You are correct in saying that this should be documented more clearly. A documentation CAR will be created to make this more clear, thanks for pointing this out.

-Ben"

-root

[BCho]

Hey Root Canal-

After further investigation into the issue and as per your request, the issue was documented as a 'regular' CAR. As Stephen has mentioned, the CAR was filed against the performance of the VI and not its documentation. I apologize for the confusion.

-Ben 

[Intaris]

As it turns out it is a normal CAR but....

We here in the forums are not privvy to private email correspondance so it's generally a good idea to

1) not perform discussions which entail knowledge of undisclosed conversations

2) inform the rest of us as to what's going on

Otherwise we get drawn into conversations where we don't belong.

Shane.

[Root_Canal]

ok, thanks. When you say "VI" do you mean that the CAR is just for the one VI I gave as an example in the original post or all of the VIs that have this bug (at least three, probably more)?

-root

[Root_Canal]

Shane, there was no email correspondence. StarsHockey's reply (above your reply) contains a hyperlink to the original post where the CAR was assigned. You can follow that link to the original post if you so desire. uvqrupkhhkpiinwg

-root