Setup your raspberry pi with Raspberry Pi instructions

Enable your camera

• Make sure your camera connected to your Raspberry Pi :

  sudo raspi-config

  select interface options and enable your camera. It will ask you to reboot if not reboot device with:

  sudo reboot

Clone this repo :

• Before clone install git:

  sudo apt update

  sudo apt install git

  git clone https://github.com/emingenc/pi-cctv.git

  cd pi-cctv

Install requirements:

• before installinng requirements.txt create venv:

  4a. Install python3-venv:

  sudo apt-get install python3-venv

  4b. Create venv:

  python3 -m venv venv

  4c. Activate venv:

  source venv/bin/activate

  4d. Install requirements.txt:

  pip3 install -r requirements.txt

Auto-start at boot:

• open etc/profile:

sudo nano /etc/profile

• add these commands end of the file

source /home/pi/pi-cctv/venv/bin/activate

cd /home/pi/pi-cctv

python3 /home/pi/pi-cctv/main.py &

 a. ctrl+x 
 b. save(press y)
 c. press Enter to save file 

apt-get requirements for image compression:

• install:

  sudo apt-get install libopenjp2-7 libtiff5

Torch and torchvision install to pi:

• you can clone this repo and install via wheel:

  git clone https://github.com/Kashu7100/pytorch-armv7l

  cd pytorch-armv7l

  pip3 install torch-1.7.0a0-cp37-cp37m-linux_armv7l.whl

  pip3 install torchvision-0.8.0a0+45f960c-cp37-cp37m-linux_armv7l.whll

  note detect.py will not run unless you dont comment out torch and torchvision from requiremnts.txt.

Train a model with dataset Check this tutorial

After you trained a model copy model weights which extension is .pt to this repo.

Change weight parameter with weight you just copied in main.py detected_objects detect function

run your main.py with object class name that you want to send post request to the server.

python3 camera.py --file-name <filename> --obj <obj_name>