I created and documented how to build the infrastructure necessary to leverage your Meraki Cameras in a fully automated “Computer Vision” Project.

To achieve this, I deployed a 3-tier application with:

• Images to work with, ideally a constant flow of images to make use of Automation.

i.e: Meraki MV 2nd gen cameras (any model ending in “2”) — they support the MV API, required for API calls.

• A relay server to capture the images when an action is triggered and forward images

i.e: Linux server in AWS to run the Python Script orchestrating all the API calls to Meraki and AWS. This server will also be our MQTT and Node-RED server running our Admin interface.

• A Computer Vision solution that will apply AI to categorize the image.

i.e: AWS Rekognition

This is a simplified view of how these will be linked:

 

 

This is what I could get from the Camera feed at the end:

Pretty great! 

Link to full article I wrote about it:  bit.ly/3d5Di6E