Accurately counting people in crowded spaces is crucial for ensuring security. It helps manage crowd movement, triggers alerts when necessary to control the flow of people in public buildings, and provides more precise attendance figures for large events.
Environments where this data is particularly valuable include train stations, airports, festivals, concerts, stadiums, demonstrations, as well as museums and large shopping malls.
Here, we present three technologies that can effectively count people in such settings, outlining their advantages and disadvantages.
Lidar (Light Detection And Ranging)
LiDAR systems work by emitting laser pulses that bounce off objects and return to the sensor. By measuring the time it takes for the pulses to return, the system generates a detailed 3D map of the environment, regardless of lighting conditions.
The map looks like a dotted maps of people and objects, as follows.
Although LiDAR sensors are widely used in autonomous vehicles, their adoption for people counting in densely populated environments is relatively recent, mainly due to the high cost of the technology.
Use case: Monitoring security lines at airports. The estimated cost for a single sensor is approximately €50,000.
Accuracy: LiDAR providers claim an accuracy rate of 98%
Companies using this technology: Outsight (France), Beonic (Australia)
3D smart cameras
An alternative to using LiDAR sensors is deploying a large network of 3D camera sensors, with one sensor positioned approximately every 10 meters to cover the same area. These cameras are installed on the ceiling at a 90° angle, providing a "bird's-eye" view of the people below.
While 3D sensors are slightly more expensive than 2D cameras, they offer significantly higher accuracy. This is because, from the vertical perspective, people do not overlap, and depth information is included, allowing the system to account for an individual’s height, thus improving the accuracy of the count.
To cover an area similar to what a LiDAR sensor would cover, about 30 3D sensors would be needed.
Use case: The estimated cost for this setup is estimated to be around €150,000.
Accuracy: 3D smart cameras provide an accuracy rate of over 99% for an area of 200m². While this approach may be at least 1% more accurate than LiDAR, it comes at three times the cost.
Companies using this technology: Xovis (Switzerland), V-count (UK), Footfallcam (UK)
2D smart cameras
A 2D camera is the most cost-effective solution for crowd counting but also the least accurate, achieving around 95% accuracy in densely populated environments. This is because it relies on detecting heads rather than full body structures, as is possible in less crowded settings.
VizioSense falls into this third category of systems. While we acknowledge that the previous two approaches offer superior performance, we believe that VizioCrowd is well-positioned in terms of pricing to deliver optimal return on investment for customers and end-users. It is 10 times cheaper than LiDAR and 30 times cheaper than a 3D camera network.
This technology can be used in various situations, such as measuring passenger flows in train stations or airport halls, monitoring audience sizes in stadiums or concerts, and counting traffic at large intersections.
See more of our use cases : industries
Conclusion: Choosing the Right Technology for the Right Use Case
The required level of accuracy will largely determine the most suitable technology.
For security-critical environments, where precision is essential, the more accurate systems are typically preferred. However, a system with 95% accuracy can still be effective for triggering alerts and managing overcrowding.
If the goal is to precisely count the number of individuals for security reasons, LiDAR offers a good balance of affordability and accuracy compared to a large network of 3D sensors.
For tasks like estimating the number of people in a security line to decide when to open new counters, 2D cameras provide sufficient accuracy and deliver a rapid return on investment.