- Flashen von Raspberry Pi OS Lite (64bit) auf die eine SD-Karte
- Headless Installation laut https://www.tomshardware.com/reviews/raspberry-pi-headless-setup-how-to,6028.html
- Verbinden mittels ssh
ssh pi@raspberrypimit dem passwort "raspberry" - Update des OS
sudo apt update && sudo apt dist-upgrade - Installation zusätzlicher Packages
sudo apt install git python3-pip ffmpeg libsm6 libxext6 -y - Klonen des repos
git clone <url> - Installation wittypy
wget http://www.uugear.com/repo/WittyPi3/install.sh && sudo sh install.shanschließend aus- und einstecken. - Installation der Pythondependencies
cd MasterArbeitCode/Basestation/raspberry_pi && pip install -r requirements.txt - Download tensorflow wheel file
wget --load-cookies /tmp/cookies.txt "https://docs.google.com/uc?export=download&confirm=$(wget --quiet --save-cookies /tmp/cookies.txt --keep-session-cookies --no-check-certificate 'https://docs.google.com/uc?export=download&id=1YpxNubmEL_4EgTrVMu-kYyzAbtyLis29' -O- | sed -rn 's/.*confirm=([0-9A-Za-z_]+).*/\1\n/p')&id=1YpxNubmEL_4EgTrVMu-kYyzAbtyLis29" -O tensorflow-2.8.0-cp39-cp39-linux_aarch64.whl && rm -rf /tmp/cookies.txt - Installation tensorflow
pip install tensorflow-2.8.0-cp39-cp39-linux_aarch64.whl - Download des weights-Files
wget https://github.com/Bleialf/MasterArbeitFiles/raw/main/yolov4-cars.tflite - Starten der Basestation mittels
python server.py <weightsfile> <wittipyfolder>mehr Informationen mittelspython server.py -h - Beispiel
python server.py yolov4-cars.tflite ../../../wittypi/ --bootdelay 100 --initdelay 100 --sleepdelay 100
Bleialf/MasterArbeitCode
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|