|
|
|
 This is a high-energy sporting event, not a piano recital. :) - Alan Anderson [more] |
 |
|
|||||||
|
||||||||
Here is my vision tracking program, so far. I can help you code anything this powerful. Just shoot me a PM. I am using OpenCV for the main libraries and using findContours and approxPolyDP to find the rectangles. This returns a boolean whether or not it can see a hot target so it can be used for the hot target detection. It is cross platform and currently runs on Windows with the Driver Station! However, It'll work in Linux, on the DS and onboard the robot because the code is very modular!
Discussion
view entire thread
16-02-2014 22:47
faust1706
Re: pic: My Vision Software, so far :)
I still do not understand why you use minenclosingcircle.....
17-02-2014 00:06
yash101
Re: pic: My Vision Software, so far :)
I didn't. I just used the "Circle" function in OpenCV's Huihgui module 
.
If anyone's interested in learning how to do this, PM me!
17-02-2014 13:16
yash101
Re: pic: My Vision Software, so far :)
Thanks! It's just an interface, running at a low resolution (320 by 240). Now, it needs to be implemented onto the robot and a connector to the cRIO needs ot be build!
18-02-2014 00:19
daniel_dsouza
Re: pic: My Vision Software, so far :)
I like how your halogen lights do not interfere with tracking the retro-reflective tape.
18-02-2014 08:36
yash101
Re: pic: My Vision Software, so far :)
I am doing a ratio test, so if you look into the HSV image, you'll see that it actually sees it and thinks it is a target. It is filtered out later and deemed as crap. By the way, I am working on getting git set up and I will make the code available after this season! 
view entire thread
Tags
loading ...