I have implimented a networked person detection system in the networkedDetection branch.
as of right now the raspberry pi can run object detection in this way at about 30 FPS, which his more than enough.
I do think it's kind of lame to have to have an extra coputer and a wifi network to run the turret, so in the future i will probably explore trying to detect people without the use of a neural net. perhaps i could do it with a color mask or something.
The mechanical side of things is still in the design phase
once a physical design is chosen, motors are chosen, and an electrical layout is chosen, code for controlling motors and such will be developed
these is just the verions i use, i don't know how far back one could go before it breaks the code
these are installed automatically by install.sh
python3-opencv=4.5.1+dfsg-5
mpv=0.32.0-3
This repo is made to be installed on a raspberry pi running raspberry pi OS or some other ubuntu based operating system
clone the repository
git clone https://github.com/Scott-McKool/turret.git
enter the repository
cd turret
make the install script executable
chmod +x install.sh
run the install script
./install.sh
to have turret.py run on startup
make the runOnStartUp.sh script executable
chmod +x runOnStartUp.sh
run the runOnStartUp.sh script
./runOnStartUp.sh
if you no longer want turret.py to run on startup, use
sudo systemctl disable turret.service