A Summer of Sensors – Part 2

In my last post, I published new work by our 2019 summer research team. Maddie and Anna have recently completed their summer with Proving Ground and I am happy to provide this new prototype that is a continuation of research into uses of the Raspberry Pi. In this study, the team created a scenario that leverages OpenCV with a PiCam to perform real-time object detection. We were interested in the potential to use a simple sensor setup to monitor human activity in a space.


Object Detection Prototype

OpenCV is an open source computer vision and machine learning library. The team set up OpenCV to run on a Raspberry Pi B+ and used a PiCam to read a video feed. After some configuration and custom scripts, the team was able to create a basic video feed setup that would report counts of people to MySQL database. From a design and operations perspective, this type of workflow has great potential to support real-time occupancy tracking of spaces – all for a hardware and software cost of under $50

While this type of setup has enormous potential to better understand our built environment, but we do need to be cautious of the the power and affordability of this technology. It is shockingly easy and cheap for anyone to potentially create a space monitoring device that could violate individual privacy or use the data for nefarious purposes. Building design, construction, and operation professionals need to be aware of the benefits and dangers of these tools.

Even still, we hope this prototype can be used to further positive research and interest in this area.

Download the Guide Here!

Sample occupancy data captured to a database over time. The graph shows counts of people in a space as the enter the field of vision for the camera.

If you are interested in hearing about Anna and Maddie’s thoughts about their work this summer, be sure to check out this episode of our Prove It video podcast on YouTube!