Creating the GRIP pipeline - GirlsOfSteelRobotics/GRIPonGalileo GitHub Wiki
Creating the GRIP pipeline
GRIP is well documented, with guides available from both FIRST and WPI:
- 2017 FRC Control System / Vision Processing / GRIP / Introduction to GRIP
- 2017 FRC Control System / Vision Processing / GRIP / Processing images from the 2017 FRC Game
- GRIP README.md
- GRIP source code on GitHub (The ultimate documentation)
To get started, experiment with a GRIP vision processing pipeline:
- Install GRIP on a desktop or laptop computer
- Download previously captured images to process with GRIP
- 2017 sample images from FIRST
- Use the GRIP documentation mentioned above to create an initial vision processing pipeline
- The last steps should be to find or filter contours
The next step is to test your pipeline with live input from a webcam:
- Complete the Intel Galileo setup steps
- Connect an ethernet cable between your laptop and the Galileo and power it up
- Attach a webcam to a USB port on the Galileo
- Mount a light ring or flashlight next to the webcam
- Log into the Galileo
- Start the video streaming server in the background
- mjpg-streamer ... &
- On the laptop, switch your GRIP input to the video stream
- Type: Network camera
- URL: http://galileo.local:1181/?action=stream
- On the Galileo, adjust the webcam exposure until the target stands out but isn't washed out to a solid white
- v4l2-ctl -L
- v4l2-ctl -c ...
- v4l2-ctl -c ... 156
When the pipeline is working correctly, export automatically-generated C++ source code:
- On the GRIP "Tools" menu, choose "?"