The following table contains the requirements for cameras to enable the effective operation of the visitor counter:

Camera

  • Resolution: 720 х 576 (CIF4) or 360 х 288 (CIF1) pixel resolution. Using pixel resolutions higher than CIF4 do not lead to higher recognition accuracy.
  • Frames per second: 25.
  • Color: color camera is obligatory.
  • No camera jitter is allowed.

Illumination:

  • Best recognition results are achieved under moderate illumination. If the scene is under- or over-illuminated, the recognition accuracy may drop down.
  • Sharp changes in illumination may lead to improper operation of analytics.

Scene and viewing angle:

  • Vertically downward position is the best for the purpose.  The closer to vertical, the more accurate counting.
  • Camera FOV dimensions: min. 2 x 2m, optimal 4 x 4m.
  • The background must be primarily static and not undergo sudden changes.
  • The counting area must not contain any moving objects except for humans.
  • Reflective surfaces and harsh shadows from moving objects can affect the quality of analytics.
  • Leafage, TV screens or any periodic object movement in the background may cause analytics glitches.
  • If possible, avoid obstruction of the humans by static objects such as pillars, trees, etc.

Images of objects within the scene:

  • Image quality: the image must be clear and sharp with no visible compression artifacts.
  • Dimensions of a human in scene: the bounding rectangle must occupy 10–60 percent of the frame area.

Other requirements:

  • The visitors must not move in a continuous flow; smaller groups of humans are counted correctly.
  • No labels