Help with using vision processing software with Java API
 

My team is in its second year and this will be the first year we are incorporating vision processing in our code for the robot. We used Java our rookie season and I am familiar with most aspects of the Java API. What I want to know is the basic steps to use a vision processing software (ie RoboRealm, OpenCV, SimpleCV), in conjuction with the Java API.

Here are my main questions:

1. Is the vision processing software running on the driver station pc or the cRIO?
2. How do you utilize data from the vision software in the actual robot code?
3. If the vision software is on the cRIO, do we need a special version of the software compatible with the cRIO and/or will it take up a lot of processing power?
4. Is it possible to use the axis camera over the network to provide a target tracking system viewable from the driver station during teleop?

Thank you so much in advance for helping me out with these issues, I am looking forward to learning a lot of cool things this season!

Re: Help with using vision processing software with Java API
 

You can choose. RoboRealm is designed for Windows, and so needs to run on the dashboard. OpenCV or SimpleCV could theoretically be run on the cRIO, but that would be a lot of work so you'd probably want to run it on the dashboard or a co-processor. NI Vision runs on the cRIO, and there is a tutorial for using it to find the vision targets. http://wpilib.screenstepslive.com/s/...95-c-java-code

We write the vision software to determine the distance to the goal and the angle to the goal. We can then use that data to turn towards the goal and drive to the correct distance
The cRIO is too slow for running high frame rate vision processing, but you may not need to process an image very often.

Yes. If this is your goal, you probably want to do the vision processing on the driver station.

Thanks for the help.