On this page you will find instructions to install a standalone camera.
Everything you need is installed on a single Raspberry Pi device:
- the camera itself
- the scanning software powered by OpenCV
- the InfluxDB database
- the Grafana web dashboard
With this architecture, use the network address of the Raspberry Pi to access live data in Grafana.
Connect over SSH to the target Raspberry Pi and apply following commands:
sudo apt-get install -y python-dev mysql-server libmysqlclient-dev libopencv-dev python-opencv
sudo mkdir -p /opt
cd /opt
sudo git clone https://github.com/bernard357/smart-video-counter.git
sudo chmod 0777 -R /opt/smart-video-counter
cd smart-video-counter
pip install -r requirements.txt
cd source
You will immediately change the id of your camera to ensure that it is unique:
sudo nano config.py
The camera id will be transmitted with counter values, and appear in the database and reports. You should avoid spaces and special punctuation characters.
Save changes with Ctl-O and exit the editor with Ctl-X.
The updater service handles measurements coming from the camera. Multiple options can be considered, for example: save data in a text file, send data to a sql database, or to an InfluxDB database.
Install the updater service with following commands:
sudo cp ../scripts/updater.service.example /etc/systemd/system/updater.service
sudo chmod 664 /etc/systemd/system/updater.service
Now you can enable the new service, start it, and get feedback:
sudo systemctl enable /etc/systemd/system/updater.service
sudo systemctl start updater.service
sudo systemctl status updater.service
If everything went fine, the system should report a successful start:
The service can be tested by sending some fake measurements, like this:
sudo python test_updater.py
The actual id of the camera should be listed. When enough samples have been
thrown, hit Ctl-C to interrupt the process. Then check the log and ensure that
samples get there.
Congratulations! At this stage your system can handle measurements in the background. Now we will setup the camera counter itself.
The scanner is powered by OpenCV. It is the sub-system that analyses in real-time the flow of images coming from the camera and generates counters every 2 seconds: number of persons, number of moves, and number of human faces detectd.
Test the scanner from the command line:
python smart-video-counter.py
Stand in front of the camera at a distance of at least 2 meters, and then get closer and show your face to the camera. The first number should reflect the number of persons, and the third one the number of faces. In the example below the camera detected that we were two persons queuing, then I moved and presented my face.
Install the scanner service with following commands:
sudo cp ../scripts/smart-video-counter.service.example /etc/systemd/system/smart-video-counter.service
sudo chmod 664 /etc/systemd/system/smart-video-counter.service
Now you can enable the new service, start it, and get feedback:
sudo systemctl enable /etc/systemd/system/smart-video-counter.service
sudo systemctl start smart-video-counter.service
sudo systemctl status smart-video-counter.service
If everything went fine, the system should report a successful start:
At this stage, the camera is scanning images, and reporting measurements in a log file. Look in real-time at the log file while standing in front of the camera:
Great! The camera is working as expected. Now we can change the configuration and put data in a real database.
InfluxDB is an open source database written in Go specifically to handle time series data with high availability and high performance requirements.
We install an InfluxDB database on the Raspberry Pi itself with following commands:
sudo su
adduser influxdb
wget https://dl.influxdata.com/influxdb/releases/influxdb-1.1.1_linux_armhf.tar.gz
tar xvfz influxdb-1.1.1_linux_armhf.tar.gz
rm influxdb-1.1.1_linux_armhf.tar.gz
cp -rp influxdb-1.1.1-1/* /
chown -R influxdb:influxdb /var/lib/influxdb
chown -R influxdb:influxdb /var/log/influxdb
cp -n /usr/lib/influxdb/scripts/influxdb.service /etc/systemd/system/
systemctl enable influxdb.service
cp -n /usr/lib/influxdb/scripts/init.sh /etc/init.d/influxdb
chmod +x /etc/init.d/influxdb
insserv influxdb
service influxd start
service influxd status
If everything went fine, you could get positive feedback on the InfluxDB service:
We change the configuration of the system so that updates are sent to InfluxDB too. Uncomment the block referring to infludb.
nano config.py
Save changes with Ctl-O and exit with Ctl-X. Then restart the updater so that the new configuration is used:
systemctl restart updater.service
After some time we can verify the presence of data in the InfluxDB database:
influx
> show databases
> use smart-video-counter
> show series
> select * from "smart-counter"
If you have multiple lines of text, you can congratulate yourself!
> exit
Grafana is an open source metric analytics & visualization suite. It is most commonly used for visualizing time series data for infrastructure and application analytics but many use it in other domains including industrial sensors, home automation, weather, and process control.
Currently there are multiple steps to install Grafana on Raspberry Pi, so let do that progressively:
# grafana data is a dependancy for grafana
sudo apt-get install -y fonts-font-awesome libjs-angularjs libjs-twitter-bootstrap
wget http://ftp.us.debian.org/debian/pool/main/g/grafana/grafana-data_2.6.0+dfsg-3_all.deb
sudo dpkg -i grafana-data_2.6.0+dfsg-3_all.deb
sudo apt-get install -f
sudo apt-get install -y golang-go golang-go-linux-arm golang-go.tools golang-src
wget http://ftp.us.debian.org/debian/pool/main/g/grafana/grafana_2.6.0+dfsg-3_armhf.deb
sudo dpkg -i grafana_2.6.0+dfsg-3_armhf.deb
sudo apt-get install -f
sudo systemctl start grafana-server.service
sudo systemctl status grafana-server.service
On successful start of the server, open a browser window and enter
the IP address used for SSH communications with the Raspberry Pi, followed by :3000.
http://<ip_of_raspberry>:3000
This will display the login page of Grafana. From there you can authenticate with admin and admin then add a new data source and build a dashboard.
Click on the Grafana logo and then select Data Sources. Select a source of type InfluxDB and then pick up the database named smart-video-counter.
Then add a new dashboard and add one graph per measure: standing, moves, faces.
