Getting Insights with Meraki Cameras and Sensors

This real showcase from the Cisco Frankfurt office highlights the possibilities of the Meraki MV cameras and Meraki MT sensors.

Currently there are 2 use-cases in one Python application implemented:

Person detection within pre-defined zones: If one or more persons are standing within the pre-defined zone (setting in the Meraki-dashboard), the camera is sending out an MQTT message to the python script. Then, only if the person stays for at least x seconds, the data (how many persons, what timeframe, for how long) will be stored in a time-series database.
Open/Close MT20 Sensor + Meraki Camera Snapshot: The webapp shows who has opened the door. It requests the last x events from the Open/Close Meraki MT20 sensor and downloads snapshots from the time where the sensor has been triggered (=door was opened) from a Meraki MV camera.

Architecture

The script uses several components:

Meraki MT sensors + Meraki MV camera 2nd generation
MQTT broker Mosquitto: to send and receive the object detection messages from the MV cameras
Time series database InfluxDB 2.0: To store the information (this script is not compatible with InfluxDB 1.x!)
Grafana for visualisation
Python
- Flask web-framework
- Paho-MQTT client

Setup your Meraki equipment accordingly, define zones in the camera settings
Install & setup Mosquitto, InfluxDB 2.0, Grafana on a Linux system
Clone this repo and deploy the python files in your virtual environment. I would recommend to use pipenv (Pipfile is included).
Edit the config.pyfile and insert your credentials and modify you configuration.
Setup your Grafana dashboard according to your zones and data.
Start the script and let it run.

1.0 - inital features: person detection with zones and open/close snapshot feature

This project is licensed under the MIT license - see the LICENSE.md file for details.