Vision Processing & Smart Dashboard help

[wildbot3844]

Our Team has decided to do vision processing for the 2014 Robotics season, can anyone give us any pointers to get us started??

And Also our Java SmartDashboard will not show our camera feed, can anyone help us there also?

[Domenic Rodriguez]

Team 341 Miss Daisy had an outstanding SmartDashboard-based vision system for the 2012 season, which they graciously published here: http://www.chiefdelphi.com/media/papers/2676?. This would probably be a good place to start.

As for the camera in SmartDashboard, did you use the standalone SmartDashboard vision installer? Without this the SmartDashboard will be missing the camera widget. Also, make sure you have enabled anonymous viewing in the Axis camera settings. See this page for instructions: http://wpilib.screenstepslive.com/s/3120/m/7932/l/93058-smart-dashboard-standalone-vision-installer

[TravSatEE]

FIRST Team 2035 posted their vision processing work earlier this year in the papers (http://www.chiefdelphi.com/media/papers/2854) collection that was used for the 2013 game. The paper's abstract is:

This paper outlines a vision processing algorithm used during the 2013 FIRST Robotics Competition game Ultimate Ascent by FIRST Team 2035. The vision processing was performed off robot on the driver station laptop and relevant information was provided to the robot over the arena network, which improved the robot response time considerably when compared to processing on the on-board cRIO. This vision processing scheme won the Innovation in Control Award sponsored by Rockwell Automation at the Silicon Valley Regional. Once the robot sees the field targets, it estimates the distance to the target. A Frisbee shooter controller can then estimate the Frisbee flight path and choose shooter aim angle to improve the chance of gaining points.

The vision algorithm worked very well, but the students did not have time to finish their full controller implementation to support automated Frisbee trajectory selection. The work was entirely done by students. The student that did the majority of the LabView work did not want to post the .vi files, but does give the full screenshot to reproduce the entire work.