| title | Vision |
|---|---|
| layout | page |
- team 900 zebracorn
- hardware:
- StereoLabs’ ZED Stereoscopic Camera
- Custom 3D printed green LED "ring" case enclosure for camera and Buck Boost Converter to eliminate flicker
- nVidia Jetson TX2 connected via USB
- NavX-MXP Gyro
- ROS w/ C++
- images down sampled by 1/2 from 720p to reduce computational overhead
- hardware:
- tuning of exposure and brightness to minimum settings to isolate for reflected bright LED light
- use of navx clock for timestamps shared between co-processor and roborio (avoid time sync)
- zebracorns 2017VisionCode/zebROS ws/src/navx publisher/
- Coordinate translation between frames of reference
- robot center x,y,z
- camera center x,y,z
- mounting point x,y,z
- odometry: where robot has moved since last initialized
- map: track robot's pose with respect to global map
- Translate target location from image based x, y, and calculated depth to real-world values:
- distance from target
- sideways (mis)alignment of target in relation to robot
- vertical (mis)alignment of target in relation to robot
- angle robot needs to turn for on target heading
- timestamp point-in-time at which preceding data was accurate for latency compensation
- 2019
- Practical Guide to State-space Control, Tyler Veness
- FRC Programming Done Right
- 2018
- 2017
- 2016
- Zebravision 4.0: Tracking
- Zebravision 4.0: Goal Detection
- Zebravision 4.0: Image Capturing
- Zebravision 4.0: Neural Networking
- Integrating Computer Vision with Motion Control, Team 254
- Taking Control of your Robot with Jared
- 2015
- Motion Planning and Control in FRC