Need help fine tuning vision algorithm

I want to get the outer 4 corners of the targets, but all the methods I’ve tried are not as accurate as I’d like. Anyone have any ideas as to how I can do that? Image of a few of my attempts

It looks to me like you have the outer edges. Have you tried to solve for the line intersections?

Greg McKaskle

Hi,

I am not sure what is the vision library you are using. If you are using OpenCV, I could suggest using the function: convexHull.

http://docs.opencv.org/2.4/doc/tutorials/imgproc/shapedescriptors/hull/hull.html

Nice work!

For more accuracy you might try using the Find Edge function (in LabView it’s under the Machine Vision palette). Once you have a rough idea of the size and location of the targets, Find Edges can do great things. I’m sure OpenCV can do something similar.

Can you post your code?