camera-sense

AI Camera with scene understanding.

Hardware Architecture

This application is based on the following hardware components:

• Raspberry Pi Zero 2 W (Wireless)
• Raspberry Pi Camera Module 3
• Raspberry Pi Zero 2 W Enclosure (requires filing of camera hole to fit Camera Module 3)
• VEEKTOMX Mini Power Bank 10000mAh

Software Architecture

The web server is based on FastAPI/Python3, while the client web application is based on Quasar/Vue/Node/Javascript. The web application allows the remote control of the camera for

• Initial camera setup of 1 picture every 10s, looping
• Timelapse photography with custom defined interval, duration and naming
• Identification of Level indicators to analyse using April Tags (todo)
• Identification of areas to monitor for change (todo)
• Alerting the client app of Level Reached, Scene Change (todo)
• Autofocus on april tags within scene (todo)
• Control of camera settings (todo)
• Control of AI conditions to monitor (todo)

Development Environment

The Raspberry Pi Zero 2 W can be used for the whole development and runtime environemnts. VS Code (Remote) can be used to easily develop the api and app within the embedded platform. The master makefile provides the following targets

• make dev-server - Start the server api and monitor for hot changes. http://localhost:8000/api and http://localhost:8000/docs
• make dev-client - Start the web client app and monitor for hot change. http://localhost:9000/
• make build-client - Build and package the web client app into dist/spa.
• make run-server- Start the production server and host the api and web client. http://localhost:8000/
• make stop - Stop the production or development server and any python3 app

Instalation

A step-by-step tutorial for setting up a Raspberry Pi Zero W 2 computer as a vision system

• sudo apt-get upgrade -y
• sudo apt-get install git -y
• sudo apt-get install cmake -y
• sudo apt-get install libssl-dev
• sudo apt-get install python3-pip -y
• sudo apt-get install python3-opencv
• pip install apriltag
• pip install fastapi
• pip install schedule
• pip install glob

Node installation

• wget -qO- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.7/install.sh | bash
• nvm install 20
• npm i -g yarn@1
• yarn global add @quasar/cli
• yarn create quasar

Autostart by adding following to .bashrc

cd ~/projects/camera-sense


export NVM_DIR="$HOME/.nvm"
[ -s "$NVM_DIR/nvm.sh" ] && \. "$NVM_DIR/nvm.sh"  # This loads nvm
[ -s "$NVM_DIR/bash_completion" ] && \. "$NVM_DIR/bash_completion"  # This loads nvm bash_completion

bash -c "make run-server"

Troubleshooting

Increase the size of iwatch

• sudo nano /etc/sysctl.conf and add line at bottom fs.inotify.max_user_watches=524288

Increase the size of the Raspberry Pi Zero 2 swap space

sudo nano /etc/dphys-swapfile
CONF_SWAPSIZE=1024