This repository is developed in Fish-IoT project

https://www.tequ.fi/en/project-bank/fish-iot/

---

tequ-rpi-setup

Raspberry PI 3/4 preparation for DAQ & AI usage.

Software	Version
node.js	20.13.1
node-red	3.1.9
tfjs-node-gpu	3.13.0
pylon	7.2.1

It is optimal to have an remote desktop connection to the RPI.

Actions

1. Prepare Raspberry Pi OS SD-card or SSD with Raspberry Pi Imager

NOTE!

If you want to use an SSD for faster read/write speeds, you can connect an SSD to your device and select that instead of an SD card. The process will be completely the same otherwise.

---

https://www.raspberrypi.org/software/

32 GB SD-card/SSD is preferable, or better.

Use 64 bit version.

Preconfigure your installation by pressing CTRL+SHIFT+X after you select OS in installer.

• SSH
• Hostname
• WiFi
• Locale settings

2. Booting and configuration

Insert the SD card in the RPI or the SSD into a USB port. Then connect to the RPI via SSH with your preconfigured settings, once connected follow below steps.

sudo raspi-config

• Enable Serial port
• Enable 1-wire
• Enable I2C
• Enable Camera

Remember to reboot after enabling settings.

3. Install Node-RED

bash <(curl -sL https://raw.githubusercontent.com/node-red/linux-installers/master/deb/update-nodejs-and-nodered)

Create users and setup passwords. If you selected no during the setup, run the command below.

sudo node-red admin init

4. Configure Node-RED to autostart on boot

sudo systemctl enable nodered.service
sudo reboot

Open browser and go to http://localhost:1880 to see if Node-RED is working. If not, run:

node-red-start

5. Install Node-RED libraries that are often needed

cd ~/.node-red
npm install node-red-auth-github &&
npm install @flowfuse/node-red-dashboard && 
npm install node-red-node-base64 &&
npm install node-red-node-daemon &&
npm install node-red-node-email && 
npm install node-red-node-exif && 
npm install node-red-node-rbe &&
npm install node-red-node-smooth &&
npm install node-red-node-suncalc && 
npm install node-red-contrib-aggregator && 
npm install node-red-contrib-binary && 
npm install node-red-contrib-buffer-parser && 
npm install node-red-contrib-calc &&
npm install node-red-contrib-counter &&
npm install node-red-contrib-fs && 
npm install node-red-contrib-fs-ops && 
npm install node-red-contrib-image-info &&
npm install node-red-contrib-image-output &&
npm install node-red-contrib-moment && 
npm install node-red-contrib-multipart-stream-decoder &&
npm install node-red-contrib-multipart-stream-encoder &&
npm install node-red-contrib-browser-utils &&
npm install node-red-contrib-sunevents  

node-red-restart

6. Using a camera

To use a camera with the RPI, you'll need the libcamera library, it should come preinstalled, but if not run:

sudo apt install libcamera-apps

sudo reboot

To test that everything is working, plug in a monitor and the camera and run:

libcamera-hello

Keep in mind to use the RPI terminal, not remote desktop terminal. If you see the camera view on your monitor, then the camera is working.

7. Install & configure hardware watchdog

sudo su
echo 'dtparam=watchdog=on' >> /boot/config.txt
reboot

sudo apt update
sudo apt install watchdog

sudo su
echo 'watchdog-device = /dev/watchdog' >> /etc/watchdog.conf
echo 'watchdog-timeout = 15' >> /etc/watchdog.conf
echo 'max-load-1 = 24' >> /etc/watchdog.conf

sudo systemctl enable watchdog
sudo systemctl start watchdog
sudo systemctl status watchdog

8. Update node.js to version 20

sudo apt update
sudo apt upgrade
curl -sSL https://deb.nodesource.com/setup_20.x | sudo bash -
sudo apt-get install -y nodejs

(Optional) Install Tensorflow & Canvas for Object Detection in Node-RED

Tensorflow tjfs-node-gpu installation

cd .node-red

npm install --ignore-scripts @tensorflow/tfjs-node-gpu

npm rebuild @tensorflow/tfjs-node-gpu --build-from-source

node

var tf = require('@tensorflow/tfjs-node-gpu')

Canvas installation

sudo npm install --global node-pre-gyp

sudo apt-get install build-essential libcairo2-dev libpango1.0-dev libjpeg-dev libgif-dev librsvg2-dev

npm install canvas 

Go to https://github.com/Lapland-UAS-Tequ/tequ-api-client for example flows how to use tensorflow

(Optional) Install RTSP Simple server

cd $home

mkdir rtsp-simple-server

cd rtsp-simple-server

wget https://tequ-files.s3.eu.cloud-object-storage.appdomain.cloud/rtsp-simple-server_v0.21.2_linux_arm64v8.tar

tar -zxvf rtsp-simple-server_v0.21.2_linux_arm64v8.tar

./rtsp-simple-server

Install plug-in to publish to RTSP server from GStreamer

sudo apt install gstreamer1.0-rtsp

(Optional) Install Basler Pylon & Pypylon

Install Basler ARM64 package and Pypylon Python libraries to use Basler camera in 64-bit Raspberry PI system-

Instructions:

• https://www.baslerweb.com/fp-1666012566/media/downloads/software/pylon_software/INSTALL~9.txt
• https://www.baslerweb.com/fp-1597837316/media/downloads/documents/application_notes/AW00162902000_How_to_Build_pylon_Applications_on_Raspberry_Pi.pdf
• https://www.baslerweb.com/en/downloads/software-downloads/software-pylon-7-2-1-linux-arm-64bit-debian/
• https://www.baslerweb.com/en/downloads/software-downloads/pylon-7-2-1-linux-arm-64-bit/

Notes:

• Remember that the USB camera must be plugged into 3.0 USB port.

cd $home

mkdir basler

cd basler

wget https://tequ-files.s3.eu.cloud-object-storage.appdomain.cloud/pylon_7.2.1.25747_aarch64_debs.tar.gz

tar -xvf pylon_7.2.1.25747_aarch64_debs.tar.gz

cd ./pylon_setup

sudo apt-get install ./pylon_*.deb 

sudo /opt/pylon/share/pylon/setup-usb.sh

pip3 install pypylon

To find pylonviewer, navigate to: /opt/pylon/bin/pylonviewer Open pylonviewer with "pylon Viewer" application which is under Sound & Video category.

(Optional) Install Gstreamer plugin for Basler cameras

Instructions at Basler github:

• https://github.com/basler/gst-plugin-pylon

sudo apt remove meson ninja-build

sudo -H python3 -m pip install meson ninja --upgrade

sudo apt install libgstreamer1.0-dev libgstreamer-plugins-base1.0-dev cmake

sudo apt-get install gstreamer1.0-tools

export PYLON_ROOT=/opt/pylon

git clone https://github.com/basler/gst-plugin-pylon.git

cd gst-plugin-pylon

meson setup builddir --prefix /usr/

ninja -C builddir

ninja -C builddir test

sudo ninja -C builddir install

gst-inspect-1.0 pylonsrc

(Optional) Use GStreamer with Basler cameras

Example GStreamer pipelines:

1. Show camera video on display

gst-launch-1.0 pylonsrc ! videoconvert ! autovideosink

2. Publish H264 encoded camera video to RTSP server

gst-launch-1.0 -v pylonsrc ! video/x-bayer,framerate=10/1,width=3840,height=2160,format=rggb ! queue ! bayer2rgb ! queue ! videoscale ! video/x-raw,width=960,height=540 ! queue ! videoconvert ! queue ! v4l2h264enc ! queue ! 'video/x-h264,level=(string)4' ! queue ! rtspclientsink location=rtsp://localhost:8554/testi204

3. Publish camera video stream as JPEG stream to TCP server

gst-launch-1.0 -v pylonsrc ! video/x-bayer,framerate=10/1,width=3840,height=2160,format=rggb ! queue ! bayer2rgb ! queue ! videoscale ! video/x-raw,width=960,height=540 ! queue ! videoconvert ! queue ! jpegenc ! queue ! tcpclientsink port=50002

4. Combine 1&2&3 in single pipeline

gst-launch-1.0 -v pylonsrc ! video/x-bayer,framerate=10/1,width=1980,height=1080,format=rggb ! queue ! bayer2rgb ! tee name=t t. ! queue ! videoscale ! video/x-raw,width=960,height=540 ! queue ! videoconvert ! autovideosink t. ! queue ! videoscale ! video/x-raw,width=960,height=540 ! queue ! videoconvert ! queue ! v4l2h264enc ! queue ! 'video/x-h264,level=(string)4' ! queue ! rtspclientsink location=rtsp://localhost:8554/testi204 t. ! queue ! videoscale ! video/x-raw,width=960,height=540 ! queue ! videoconvert ! queue ! jpegenc ! queue ! tcpclientsink port=50002

(Optional) Use GStreamer with native Raspberry PI HQ camera

1. Show camera video on display

gst-launch-1.0 -v libcamerasrc ! video/x-raw,framerate=10/1,width=1920,height=1080 ! queue ! videoconvert ! autovideosink

2. Publish H264 encoded camera video to RTSP server

gst-launch-1.0 -v libcamerasrc ! video/x-raw,framerate=10/1,width=1920,height=1080,format=NV12 ! queue ! videoconvert ! queue ! v4l2h264enc ! queue ! 'video/x-h264,level=(string)4,video_bitrate=300000' ! queue ! rtspclientsink location=rtsp://localhost:8554/testi204

3. Publish camera video stream as JPEG stream to TCP server

gst-launch-1.0 -v libcamerasrc ! video/x-raw,framerate=10/1,width=1920,height=1080 ! queue ! videoscale ! video/x-raw,width=960,height=540 ! queue ! videoconvert ! queue ! jpegenc ! queue ! tcpclientsink port=50002

4. Use libcamerasrc to stream through rpicam

gst-launch-1.0 -v libcamerasrc ! video/x-raw,framerate=10/1,width=1920,height=1080 ! queue ! tee name=t t. ! queue ! videoconvert ! autovideosink t. ! queue ! videoscale ! video/x-raw,width=960,height=540 ! queue ! videoconvert ! queue ! v4l2h264enc ! queue ! 'video/x-h264,level=(string)4' ! queue ! rtspclientsink location=rtsp://localhost:8554/testi204 t. ! queue ! videoscale ! video/x-raw,width=960,height=540 ! queue ! videoconvert ! queue ! jpegenc ! queue ! tcpclientsink port=50002