A realtime vision program for FRC teams.
If you are on a macbook (school-issued in our case), run ./scripts/install_schoolmac.sh.
If you are on an Debian/Ubuntu distribution of linux, run sudo ./scripts/install_opencv_debian.sh
And then, run pip3 install -r requirements.txt
Set ./TE.AM's contents to your team number seperated into the middle IP notation.
For example, if my team number was 3966, ./TE.AM contains 39.66. If it were 235, the file would contain 2.35, etc.
Set ./R.PI to the last octet for the raspberry PI's static IP address. This is important to remember, as you will need to connect via a static IP on the field, because mDNS will not work (which is how you connect via raspberrypi.local. Don't worry, when you are off of the FMS (back at a robotics lab), you will still be able to connect via raspberrypi.local. So, if I want to connect to the raspberry pi at 10.39.66.176, I would set ./TE.AM to 39.66, and ./R.PI to 176.
Now, run ./scripts/deploy.sh
ssh to the raspberry pi (ssh [email protected]), then cd ~/, and tar xfv ~/Downloads/punkvision.tar.xz.
Now, cd punkvision, and ./scripts/install_raspi.sh.
Create some auto-starting files by copying ./scripts/auto_nothing to something that starts with auto_. i.e. ./scripts/auto_ABC.sh.
After this, run ./scripts/update_raspi.sh. You will need to do this each time you change networking information, or add or remove an autoscript.
Now, reboot, and it should run the autostart scripts on boot!
The address should be raspberrypi-3966-FRC.local
- Use cscore for streaming to the driver station (right now, you have to open a web browser)
- Provide raspi images for teams to use
- Add Makefile to install correct packages, and detect some things (probably not enough to warrant a ./configure step)
Please report bugs to the issues
Here are a few common ones:
"Could not find any downloads that satisfy the requirement opencv-python" while installing dependencies.
- To fix this, run
sudo pip3 install --upgrade pip