Machine Vision

cancel
Showing results for 
Search instead for 
Did you mean: 

display the detected image

Hi,

 

I don't quite understand what are you trying to do... Could you please explain the nature of your application?

 

So, you create a screenshot image of your program (Artemis ATK) with the image. Doesn't this application have the capability of saving the image only? You only need the white shape with the black background? Is this your only image, or do you have a set of such images?

If this is a single image, use a simple program (like Paint) to crop the image. You can of course do this in Labview also using the extract image functionality. If you have a set of such images you can manually select the region of interest for only first image and then loop through all other images (only if the position of the window remains the same). If you need to extract the image with subpixel accuracy, you could possibly search for edges along all four sides.

 

Again, I really do not know what are you trying to do. Maybe someone else better understands what are you trying to achieve.

Regards,

K


https://decibel.ni.com/content/blogs/kl3m3n



"Kudos: Users may give one another Kudos on the forums for posts that they found particularly helpful or insightful."
0 Kudos
Message 11 of 16
(1,171 Views)

Hello Klemen,

Actually I am working of electroluminscence imaging process and this is a sample picture. Yes, I do have set of picutre but they are all the same way and I just need to display the white area of the picture which is actual cell. I am attaching some more picture. Could be this one can explain better. 

thanks 

 

regards

Download All
0 Kudos
Message 12 of 16
(1,165 Views)
Hi,

maybe you could try the rotation invariant pattern matching approach. Or geometric matching if you have changes in scale throughout your images. Take a look at the Ni Vision examples to see if this approach suits you...

Regards,
K

https://decibel.ni.com/content/blogs/kl3m3n



"Kudos: Users may give one another Kudos on the forums for posts that they found particularly helpful or insightful."
0 Kudos
Message 13 of 16
(1,161 Views)

Hi, 

 

Thanks for this advise. I will look at this and come back to forum if stuck somewhere. Thanks for all help. 

 

regards

0 Kudos
Message 14 of 16
(1,156 Views)

To correct my previous post - i saw your cell images differ from one another (was on my cell phone and missed them somehow - my fault). So the correlation (pattern, geometric matching) might not be the best approach.

 

Try to take a look at the thresholding and morphology examples too.

 

Also, I noticed that your cells are rotated. You cannot (to my knowledge) display only the rotated parts (like I've said, the images are represented as arrays with rows and columns). You will get some padding.

 

Best regards,

K

 

 


https://decibel.ni.com/content/blogs/kl3m3n



"Kudos: Users may give one another Kudos on the forums for posts that they found particularly helpful or insightful."
0 Kudos
Message 15 of 16
(1,146 Views)

Hello Klemen, 

 

I am trying to get something out of your kind advices if I stuck and couldn't do it I will comback to the forum for help but I really want to spend two or three days myself and want to see what outcome should I manage to produce. 

 

regards

0 Kudos
Message 16 of 16
(1,126 Views)