A Camera for the Raspberry Pi

Raspberry Pi + Camera

It is time to have a companion for my Raspberry Pi 4.

Hardware

• Arducam OV5647 (5MP, 1080p) -

• Pi Zero Wifi Security Camera - https://www.youtube.com/watch?v=ll5d342QaCY with MotionEyeOS
  • https://randomnerdtutorials.com/install-motioneyeos-on-raspberry-pi-surveillance-camera-system/

Taking Pictures and Streaming Video

ssh re@192.168.0.232

Enable the camera:

sudo raspi-config #Interface Options
#-->> I1 Enable (legacy) camera Support #bullseye 6.1kernel
sudo reboot #you will be prompt to reboot

Take a picture simply with:

raspistill -o pico.jpg
#scp yourraspbiuser@192.168.0.232:/home/reisipi/pico.jpg . #copy it to windows CMD

Save Pictures Every 60s

Get the sample script at my RPi Repo!

mkdir -p ~/growth
nano capture_images.sh
./capture_images.sh

Pi Camera - Video Streaming

And now video streaming with:

git clone https://github.com/meinside/rpi-mjpg-streamer 
#git clone https://github.com/JAlcocerT/rpi-mjpg-streamer
cd rpi-mjpg-streamer

Build the image (use the user/pass that you’d like!)

sudo docker build --no-cache -t streamer:latest \
    --build-arg PORT=9999 \
    --build-arg RESOLUTION=400x300 \
    --build-arg FPS=24 \
    --build-arg ANGLE=0 \
    --build-arg FLIPPED=false \
    --build-arg MIRRORED=false \
    --build-arg USERNAME=user \
    --build-arg PASSWORD=some-password \
    .

This took ~2min in my RPi4 2GB

Deploy the container image: with CLI or with Docker-Compose

docker run -p 9999:9999 --device /dev/video0 -it streamer:latest

version: '3'
services:
  streamer:
    image: streamer:latest
    container_name: streamer_rpi_cam
    ports:
      - "9999:9999"
    devices:
      - "/dev/video0:/dev/video0"
    stdin_open: true
    tty: true
    restart: unless-stopped

And we will have the interface ready at the specified port: localhost:9999 and user/password as per your specifications while building the image.

Working up to GNU Debian v11 - Bullseye (not available in bookworm!)

More Software - Camera and CV

Pi NVR

An NVR stands for Network Video Recorder. It’s a specialized computer system designed to record and store video footage captured by IP (Internet Protocol) cameras.

• https://github.com/geerlingguy/pi-nvr

Raspberry Pi NVR for home CCTV recording.

TensorFlow Lite

• A)

git clone https://github.com/tensorflow/examples --depth 1 #https://github.com/JAlcocerT/tf-examples
cd examples

Create a Py venv:

python3 -m pip install virtualenv

# Successfully installed distlib-0.3.8 filelock-3.15.4 platformdirs-4.2.2 virtualenv-20.26.3

python3 -m venv tf
source tf/bin/activate
cd ./lite/examples/object_detection/raspberry_pi
sh setup.sh

Or a Conda one if you prefer:

wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-armv7l.sh

chmod +x Miniconda3-latest-Linux-armv7l.sh
./Miniconda3-latest-Linux-armv7l.sh
source ~/.bashrc
#conda --version

conda create --name myenv python=3.7.3
conda activate myenv
python --version

Once all dependencies are installed in the chose isolated environment:

python3 detect.py --model efficientdet_lite0.tflite

• B) Raspberry Pi Object Detection Tutorial

OpenCV

PyTorch

Yolo (ultralytics)

With Yolo Models - https://github.com/ultralytics/yolov5

MotionEyeOS

A Video Surveillance OS For Single-board Computers

• https://github.com/motioneye-project/motioneyeos
  • https://github.com/motioneye-project/motioneye/wiki/Install-In-Docker#image-from-docker-hub

docker run --name="motioneye" \
    -p 8765:8765 \
    --hostname="motioneye" \
    -v /etc/localtime:/etc/localtime:ro \
    -v /etc/motioneye:/etc/motioneye \
    -v /var/lib/motioneye:/var/lib/motioneye \
    --restart="always" \
    --detach=true \
	ccrisan/motioneye:master-armhf
    #ccrisan/motioneye:master-amd64 #https://hub.docker.com/r/ccrisan/motioneye

Now, you will access MotionEye at localhost:8765 with default user: admin and blank (no password).

motionEyeOS is a Linux distribution that turns a single-board computer into a video surveillance system.

Frigate

• https://docs.frigate.video/

A complete and local NVR designed for Home Assistant with AI object detection. Uses OpenCV and Tensorflow to perform realtime object detection locally for IP cameras.

NVR with realtime local object detection for IP cameras

Other Options for CV with RPi

• https://github.com/silvanmelchior/RPi_Cam_Web_Interface
• https://www.youtube.com/watch?v=tT5gHNDBHXo

FAQ

• HA Desktop - https://flathub.org/apps/com.cassidyjames.butler

MJPG-streamer takes JPGs from Linux-UVC compatible webcams, filesystem or other input plugins and streams them as M-JPEG via HTTP to webbrowsers, VLC and other software.

It is the successor of uvc-streamer, a Linux-UVC streaming application with Pan/Tilt

• https://github.com/jacksonliam/mjpg-streamer
  • https://sourceforge.net/projects/mjpg-streamer/

AI

Best Performing LLMs for a RPi

• For code: https://ollama.com/library/deepseek-coder

  • https://ollama.com/library/phind-codellama

• Daily: https://ollama.com/library/gemma

  • https://ollama.com/library/orca-mini
  • https://ollama.com/library/tinyllama

How to use Ollama with a Raspberry Pi

You can SelfHost Ollama with Docker and choose a small enough model that it will fit in the RAM.

Build multimodal AI applications with cloud-native stack:

• https://github.com/jina-ai/jina

  • https://docs.jina.ai/cloud-nativeness/opentelemetry/

• https://github.com/antimatter15/alpaca.cpp

BI Tools for EDA

• https://www.opensourcealternative.to/project/LightDashs
• https://www.opensourcealternative.to/project/Metabase

More

How to Install Raspberri Pi camera and Troubleshoot errors | Upgrade Bios Firmware on Raspberry Pi

• Node-RED with Raspberry Pi Camera (Take Photos) - https://randomnerdtutorials.com/node-red-with-raspberry-pi-camera-take-photos/

• Video Streaming with Raspberry Pi Camera - https://randomnerdtutorials.com/video-streaming-with-raspberry-pi-camera/

• Pi Zero W + Camera + MotionEyeOS (https://github.com/motioneye-project/motioneyeos) - How To Make A Raspberry Pi Zero WiFi Security Camera, Also Accessible Over The Internet - https://www.youtube.com/watch?v=ll5d342QaCY

  • https://github.com/motioneye-project/motioneyeos/wiki/Supported-Devices