The computer vision is critical for the vehicle to obtain situational awareness of it's location in the pool as well as its location relative to the obstacles.
The vehicle relies on a downward pointing camera and a forward looking camera (to be installed in Vehicle 2). Camera images are processed with the onboard Odriod computers to extract key information about the markers and obstacles in the pool. For example, orange markers are placed on the bottom of the pool to tell the vehicle the direction to the next task. (see video below) The computer vision algorithm detects these markers and determines their orientation relative to the vehicle. The orientation is then reported to the autonomy node, which then commands the vehicle controller to rotate and steer the vehicle in the direction of the orange marker.
More info coming soon.