Our 2017 vision code
Generated by GRIP $LINK, but then tweaked so that it works (TODO: Which lines)
To setup a development machine (your laptop), first clone this repo:
git clone https://github.com/LN-STEMpunks/WOPR-JR-Vision.git && cd WOPR-JR-Vision
Now, install requirements:
Run ./install.sh and be guided through the process
We don't have a tested method to get it working on these platforms.
On your roborio/raspi/jetson/whatever you will need to install what ./install.sh does.
We will include a raspi image with all this installed that you can download (coming later).
To deploy to a raspberry pi, run
./deploy.sh
You can also use:
./deploy.sh computername.local:~/path/to/vision/ [-p password] (replacing with the raspis name or IP, and where you want to put the contents of ./src)
To view help, run python src/grip.py -h
Mainly, when developing, you will use:
python src/grip.py --show
On the raspberry pi or jetson, use:
python src/grip.py -f thing.conf -ip roboRIO-XXXX-frc.local --publish
Where thing.conf is a copy of lab.conf, but tweaked for your current setting,
roboRIO-XXXX-frc.local is the network tables address
and --publish tells us to publish the findings to networktables
Essentially, this finds contours within a certain threshold which is in lab.conf which is the reflective tape.
Then, it takes the two largest contours, finds their center, and then plots a point (and publishes the coordinates to networktables) on the image