How to specify the ROI

[Brzhou]

Hi Sir, 

I am new for vision, i try to find the contour from the pic, but always report that the ROI is not correct.

Could someone give some examples for how to define this ?

Thanks.

[Walker34 · Accepted solution]

Maybe you could use NI Vision Assistant to generate a vision script. It is much easier to defines ROI or such things with that assistant. Then you can create a VI out of your script.

For your answer, Global Rectangle Array ROI Descriptor is responsible for ROI. 

From upper-left corner of image in pixels: left-edge, up-edge, right-edge, bottom-edge :

[Brzhou]

Hi Sir, 

Sorry, still report "invalid ROI", do you have any samples for capturing the contours, dots, distance/angle calculation ?

Thanks.

[Bob_Schor]

Hello, @Brzhou.

When I type "LabVIEW define Region of Interest", I get a number of "hits" (including some videos) that describe both manual (you looking at an image and drawing a circular, rectangular, or other shape on the image to define the ROI) and programmatic ways to create ROIs.  Once you have an ROI, you can use it to "isolate" an Image for further processing.

Welcome to LabVIEW Vision.  After using "ordinary" LabVIEW for about 5 years, a colleague asked me to help him with some code that used LabVIEW Vision.  It was quite a "change" from the way I had been handling data, and took the two of us struggling together to "teach each other" how to do what we were trying to do.  But this is an important point -- learning "new Tricks in LabVIEW" is much easier if you do it with a partner.  I also recommend writing little Demo programs (the simpler the better) to test out simple ideas (like putting an Image in a Display Window and creating ROIs on the Image).

Bob Schor

[Brzhou]

Hi Bob, 

Thanks for sharing your experience.

Yes, that is the similar way i have done since many years ago when i first time saw labview code.

I image some scenarios of the real test program of our testing stations, and after work, i try to realize it.

And this time, i try to image some scenarios we may have during the coming job, then try to do coding in advance, to see what kind of issues may happen.

Thanks, buddy!

[Brzhou]

I have watched some videos, although it is not so clear, haha.

Sounds like it is more convenient to use Vision Assitant to generate the VIs, but for some functions, still not so familar to it.

Thanks.