An ability to remove children from the total people count by estimating and filtering height has been promoted as one of the key features of people counting technology for several years. Now with the Irisys Vector 4D coming to market, there is, for the first time, an accurate and reliable way of actually measuring height.
Until now heights have been estimated by stereo video sensors using technology by taking two slightly different video images and using them to make a best guess at the height of the people passing underneath the sensor. This technique also relies on accurately knowing the distance between the sensor and the floor. These methods provide an estimate of height but the reliability of this measurement can be heavily influenced by changing light conditions and patterns on floors or clothing. They also depend on an accurate measurement of detector height above the floor, which requires height calibration on initial setup. Subsequent estimations of height will be inaccurate if this step is performed incorrectly. Finding an accurate height also typically requires mounting above a level floor surface.
The time-of-flight (TOF) sensing technology incorporated into the Vector 4D overcomes these problems. TOF sensing provides direct height measurement of everyone passing beneath the sensor, to a small margin of error. The Vector 4D sensor also builds up a baseline impression of the area it is installed in, automatically measuring its installation height and determining the nature of the floor underneath it. Whether this is a simple flat surface, a sloped surface or even stairs, the sensor incorporates this into the calculations of height for the people passing through the counting area.
These features mean that height data can be collected in virtually any location, that installation are faster and easier than ever before, and that sensors provide more accurate and reliable data. Once accurate data is provided it is possible to accurately decide the number of children and adults in the location – by measurement not estimation.