Brussels / 3 & 4 February 2018


Mirai and Computer Vision

Security and function of connected webcams

In this demonstrative session we learn the ropes of IoT nodes with computer vision. We develop a web camera application, consider security, and store or forward data for image processing.

Combining a number of components to make a full IoT system solves end to end use cases like secure image processing via webcams. With the limited processing power and storage of embedded systems, we learn to move data across networks, routing via gateways, and messaging to cloud applications that provide back end processing. To this end, we'll put a number of pieces of hardware in operation and take a cookie cutter approach to conclude with a system consisting of:

  • webcams
  • image sensor hosts
  • IoT gateways
  • cloud servers
  • imaging SDKs

Key learnings include:

  • Understanding of IoT optics
  • How to connect to AWS IoT via MQTT
  • Collecting video to process with OpenCV
  • Leveraging AWS Rekognition and deep learning

Hardware will be demonstrated in this typically hands on (but adapted to room conditions) session. Novice users are welcome, and we will cut and paste source code with step by step instruction maintaining a level of learning and fun.


Photo of Michael Schloh von Bennewitz Michael Schloh von Bennewitz