The Evolution of People-Counting Technologies
Northland Controls’ Rob Kay gives a primer on the different types of people-counting technologies and which applications suit them best.
From the humble origins of using human resources to manually tally the number of people entering or exiting a space, people-counting technology has come a long way. Every people-counting requirement is different, and there is no “one-size-fits-all” solution.
The correct application will be influenced by existing infrastructure, budget, accuracy requirements and privacy concerns. What follows is an overview of the types of people-counting technologies that exist.
Historically people counting heavily relied on a pair of IR beams. Installed on either the side of a door or an entranceway, beams were broken, and the direction (if multiple beams used) and volume of traffic could be calculated. This worked well under perfect environmental conditions and people behaving in an expected manor.
Unfortunately, this setup was susceptible to erroneous counting by things as simple as someone dragging a bag behind them, or people walking side by side. These are devices installed specifically for people-counting and mounted in order to achieve best results.
The next generation of people-counting technology made use of human written and configured analytics, predominantly making use of camera footage in the visual light spectrum. This applied the concepts of IR beams to live or recorded video.
One or more “trip lines” could be configured, and once again direction of travel and volume could be established. Many camera and analytics vendors brought algorithms to market in an effort to increase accuracy by determining from the video data if the objects crossing lines were people, cars, squirrels or trash blown in the wind.
Thermal cameras saw some use as they provide a different variable (source heat) as input into any analytical algorithm to determine the “humanness” of a target object.
Generally, existing cameras are used in these applications. This can create challenges as these cameras are designed and installed to provide the best coverage from a security perspective, and not the best angles for people counting.
More recent law and regulation changes, such as GDPR in Europe and CCPA in California, have raised concerns and red flags around the use of video in people-counting applications. This has led to new technologies emerging that use IR lasers for 3D input or LIDAR (radar in the light spectrum).
This allows sensors much of the data available in the visual light spectrum while still preserving privacy. These devices are generally designed and installed to achieve best results for people-counting and can provide some useful benefit from a security perspective.
The phrase AI is used a lot, so for clarity, in this instance we are referring to the use of machine learning. Human-written analytics will only ever be as good as the person or team that wrote them. Using the input from the above discussed sensors, machine learning will process data provided in order to “train” the model on what to recognize.
Initial investments are needed to provide the input data, but once this has been completed the model will be constantly updated and tweaked for the scene. This generally results in higher accuracy and a higher tolerance for environmental changes in the scene.
All of the aforementioned technologies focus on detecting the movement of people, or people-shaped objects. However, other strategies are available.
Wireless access points can be used to count the number of unique MAC address requests — in essence, counting devices like laptops or phones.
Bluetooth or RFID sensors can be deployed on a network to provide high accuracy tracking, or, for some, data held in existing access control or visitor management systems may provide the required information.
As the cost to deploy people-counting solutions continues to decrease and accuracy continues to improve, this technology is being more widely adopted. Once the domain of retail or higher security areas, people counting enables security departments to provide real-time actionable data to the rest of the business.
Space occupancy helps real estate and facilities teams understand people flow and where space needs to be increased or reduced. Daily occupancy helps catering teams and gym facilities to staff and understand the daily demands on them.
The only question left to answer is: how could data about the movement of people in your space help you, and help your business?
Rob Kay is Director of Professional Services at Northland Controls. He is also a member of the PSA Emerging Technologies Committee.
If you enjoyed this article and want to receive more valuable industry content like this, click here to sign up for our FREE digital newsletters!
Security Is Our Business, Too
For professionals who recommend, buy and install all types of electronic security equipment, a free subscription to Security Sales & Integration is like having a consultant on call. You’ll find an ideal balance of technology and business coverage, with installation tips and techniques for products and updates on how to add sales to your bottom line.
A free subscription to the #1 resource for the residential and commercial security industry will prove to be invaluable. Subscribe today!
Great article, lots of truths around the issues that plague accurate results. I’d love to hear your thoughts on some of the newer technologies and how AI & ML play into it. Your statement “Human-written analytics will only ever be as good as the person or team that wrote them…” is something I think many people don’t truly understand. If it were easy, it would have been done already.