Please create more videos like this one for Cognex Vision systems. I am attempting to teach myself this stuff, and this video was by far the best video i have found on the subject!
Thank you for the video it was great! Concerning lighting that can have a major issue on different algorithms depending on how they work like a basic threshold operation. For what is termed normalized grayscale pattern matching a change in light level does not affect it that much same with edge detection routines. Along with light level making sure your image is always in focus can also be of importance. People who design optical systems use the term blur diameter and keeping the object in focus keeps that value the same. Another possible hook is changing the camera gain may help get you back to the response you had before a lighting change happened. Adjusting a light level using computer control could possibly help depending on inspection speed.
Thank you for the video it was great! I have written a lot of image processing code (low level) such as how the pattern tool would work. I expect it is what is called normalized grayscale pattern matching which computes a correlation coefficient. For the die with 6 dots the issue is only 3 dots were trained so think of it as image subtraction and this die has 3 more dots which is going to add to the error. If the edges of the die were trained you would not have the issues of different numbers of dots, but thresholds can always be set.
Hi, Why is it When I load perfectly regular sized images, that when I open that folder up in Emulator all of my images are zoomed in on a a default setting. Even when hit zoom out, the zoomed in picture stay the same, zooming in or out? Thanks.
Also, im not sure this is even possible, but if someone could upload the zip files used in this exercise, i think that would really help someone like me.
Please create more videos like this one for Cognex Vision systems. I am attempting to teach myself this stuff, and this video was by far the best video i have found on the subject!
I really like the teacher's tone and humorous
Amazing explanation, thank you for posting the webinar.
Thank you for the video it was great!
Concerning lighting that can have a major issue on different algorithms depending on how they work like a basic threshold operation.
For what is termed normalized grayscale pattern matching a change in light level does not affect it that much same with edge detection routines.
Along with light level making sure your image is always in focus can also be of importance.
People who design optical systems use the term blur diameter and keeping the object in focus keeps that value the same.
Another possible hook is changing the camera gain may help get you back to the response you had before a lighting change happened.
Adjusting a light level using computer control could possibly help depending on inspection speed.
Thank you for the video it was great!
I have written a lot of image processing code (low level) such as how the pattern tool would work.
I expect it is what is called normalized grayscale pattern matching which computes a correlation coefficient.
For the die with 6 dots the issue is only 3 dots were trained so think of it as image subtraction and this die has 3 more dots which is going to add to the error.
If the edges of the die were trained you would not have the issues of different numbers of dots, but thresholds can always be set.
amazing .. more than amazing , I liked it so much .. Cognex!!
What a great tutorial. Thanks for sharing.
Best Vision Tutorial thanks !
Great lovely teacher!
Wow!! Great Tutorial
Hi, Why is it When I load perfectly regular sized images, that when I open that folder up in Emulator all of my images are zoomed in on a a default setting. Even when hit zoom out, the zoomed in picture stay the same, zooming in or out? Thanks.
it's because of the resolution of the camera that you using
How to use SDK developer software ?
is there any subsoftware cam-cic-5000r-14-g?
nice lection, bad mic,
Also, im not sure this is even possible, but if someone could upload the zip files used in this exercise, i think that would really help someone like me.