How people counter sensors work: A tech comparison

Closed-circuit television (CCTV) is one of the most common people counter technologies in use today, though it is primarily employed for security purposes rather than data acquisition. Retailers make pervasive use of CCTV systems to improve business security by deterring criminals and providing visual evidence in the event of criminal activity. 

Prior to the onset of intelligent video analytics and on-board computer vision, people counting with CCTV was a time-consuming manual process. Now, visual optics and simulated stereo vision can be supplemented with machine-learning for automated facial recognition and object detection. This can cost significantly more than standard video surveillance products, like CCTV.

The main drawback of these people counting systems is the lack of anonymity, since they are aimed at security purposes, the growing concern of personal privacy is leading to a backlash against this technology, also, it is prone to counting errors since their counting accuracy is limited by ambient environment factors such as low-light and shadows.

Tracking WiFi probe requests is an advanced, if inaccurate, method of counting people. Estimates suggest that there are as many as 6.92 billion smartphone users worldwide, each device having a unique Media Access Control (MAC) address. With WiFi-enabled, devices broadcast their MAC address in what is known as a probe request. Internet routers can acquire these signals and triangulate the position of the device based on signal strength. Multiple routers offer a more detailed insight into device position by comparing relative signal strengths.

There are many problems with WiFi people counters. First, they operate on the assumption that all people are carrying a smart device with WiFi-enabled. Second, positional accuracy is poor compared to most other people counting technologies. Third, it is controversial as the system tracks user devices without asking for consent.

Based on innovative Time-of-Flight (ToF) technology, infrared emitters offer an unprecedented combination of precise and private people counting performance. Using 3D IR depth sensors for counting people with an expanded Field-of-View, ToF people counters detect objects as a function of interrupted signal beams. This is used to monitor people entering/leaving a specific area in real-time with exceptional accuracy and complete anonymity.

Conventional infrared, or thermal, sensors have been criticized for poor accuracy in terms of false object detection and lack of clarity between multiple objects within a single Field-of-View. At Terabee, we have resolved these issues with the Terabee People Counting L-XL, an innovative counting solution which features an innovative and proprietary people counting algorithm, offering 98% accuracy. This is achieved with advanced filtering techniques to remove static people and objects that might otherwise reduce the counting accuracy.