Machine Vision

cancel
Showing results for 
Search instead for 
Did you mean: 

circle quality

Hello,


I'm using LabVIEW 8.2 with VDM 8.2,And I want to inspect the circle quality ,what tool can I use?
0 Kudos
Message 1 of 5
(4,042 Views)
You'd better show us the detail about your problem!
0 Kudos
Message 2 of 5
(4,027 Views)
First of all, i apologize about my english !
in the image have two bright ring, one is  orbicular,and one is only a part of ring,i want to inspect that whether  the two bright ring orbicular or not. 
any ideas?
Thank you !!
0 Kudos
Message 3 of 5
(4,002 Views)
那我说中文好了,我不知道怎么用vdm,但是vision builder倒还懂一点儿,
如果你拍摄的图片的大小是固定的,可以试一下我附件中的方法啊!
原理很简单,就是二值化后,然后填充圆环,如果内环和外环都是完整的,那么它填充后的面积应该是一个相对固定的值
可以取改值为你的检测阈值(适当设定一个误差),如果填充后面积在误差范围内,那么就认为内外环都是ok的,要么就不完整;
试一试,刚想到的,方法太笨,不知道行不行,(例子不完整,只给出了思路后面的处理部分自己添加)。
我在QQ建了一个机器视觉的交流群,欢迎加入,我的号码:247749311
0 Kudos
Message 4 of 5
(3,999 Views)
Hi Kert!

I am not sure, if I understand what exactly you want to measure.
Nonetheless, some ideas...

You could use particle analysis after thresholding the image. That puts the image into the binary domain. Pixels are either on or off. (0 or 1)
Now you could select e.g the aspect ratio for each circle to determine, how "round" each circle is.
In order to see, if they are located around the same center, you could use the center of mass for each circle and examine the difference between the center of mass (C.O.M) x and y coordinate.

I hope that helps somewhat.

With some more detailed explanations on what exactly you are trying to accomplish, we could provide you with more ideas on how to solve your problem.

Regards,
Markus Tarin

0 Kudos
Message 5 of 5
(3,973 Views)