6.2 Ensure overlap in the two detection zones
The algorithm example relies on the fact that a person crossing an area being tracked must first be detected
in the first zone, then in both zones at the same time, and finally in the second zone. In Figure 7. Example
of captured distances, both tracking paths have a few dots corresponding to the moment when a person was
detected in both zones simultaneously. Increasing N, increases this number of dots, but could decrease the
number of dots corresponding to the time when a person is detected only in the first or second zone.
To optimize the behavior of the algorithm, it is advised to choose N as small as possible and to apply a
filter, on the detected distances, which consists of considering only the minimum distance value from the last
Z measurements, up to the time where a person is detected in the last zone. This significantly optimizes the
probability of detecting a person in each of the zones of the tracking path.
The figure below shows the well-defined detection of a person in each of the three zones.
Figure 12. Filter on the measured distances with Z = 10, N = 6, M = 16
UM2600
Ensure overlap in the two detection zones
UM2600 - Rev 2
page 10/15
To Next Page
Loading...
