Vision: Tacking Disc on the Ground

[cbolinger]

Use the Camera to track a disc(s) on the ground. This will ensure that we drive straight at the two disc under the pyramid and the two in between the pyramid and the driver station.

See the following YouTube video of an example and proof of concept: http://www.youtube.com/watch?v=sL6hQ_w5y_8

We only need to track white disc.

[szechy]

The nice thing about that video is that they were tracking a blue disc. So, they were probably filtering the image by color, and then looking for the blue. If we are looking for white discs in a place that has any sort of light or other white things, it will be tougher.

For tracking straight in autonomous, I think it will be easier to use the gyro (especially if we do some sort of autonomous where we pick up the discs right in the center of the field), and vision tracking on the high goals themselves.

[Tavaru5]

So far we have a functional program to track white discs on the ground. All it needs is a little bit of tweaking of the RGB/HSL for the color recognition and tweaking of the autoadjust for auton. Other than that, the code should be operational.

[szechy]

elieve the next step is to actually have the program run inside of a Dashboard on a PC and have the cRIO read data from the Dashboard. That can be done with any functioning cRIO on any robot - especially the shooter testbed, since the shooter appears to be shelved for now.

At the same time, I will start working to incorporate that for testing for autonomous.

@cbolinger We will also need a camera mount on the practice robot for this, somehow.

[Tavaru5]

We have managed to get both sets of vision code running and recognizing images with outputs such as target pixels off and disc pixels off. It will however take a little bit of work to merge it into everything else.