Page History

Configuring the parameters

To configure the Close-standing people detector, do the following:

1. Create the Human pose detector (see Configuring the Human pose detector).
2. Below the Human pose detector, click Create… → Close-standing people detector.

...

By default, the Close-standing people detection tool triggers detector generates an event when "skeleton" bounding boxes around individuals people collide. 

If necessary, you can change the detector parameters. The list of parameters is given in the table:

To save the parameters of the detector, click the Apply Image Added button. To cancel the changes, click the Cancel Image Added button.

Configuring the frame perspective 

If you want to change the distance that must generate an event, you must configure You can set another triggering distance by adjusting the frame perspective:

1. In the preview window, click the Image Added button.
2. Specify
3. Select a detection tool, and click the Image Removed button in the preview window.
Set
4. the size of the same object in different
areas
5. parts of the frame
.
6. by eye:
   • To create a leveling rod, left-click
on
7. • the video image
and
8. • to add two anchor points.

   • Set at least

three leveling rods.

9. • three calibration rods. The lower anchor points of any three rods must not be located on the same line (horizontal, vertical, slanting, and so on).

   • The size of the leveling rod
should
10. • must be about the average height of a person at this
point in
11. • part of the frame.
    • You can resize the rod by stretching its anchor
points Image Removed
12. • points Image Added. You
can move
13. • can move it on screen
by
14. • by dragging and dropping.
      
Image Removed
15. • Image Added
      

      Note
      title Attention!
      When

setting

16. • you set up

the

17. • leveling rods, the following conditions

should

18. • must be met:

The

19. • the rods that are located at the same distance from the

video

20. • camera lens

, should

21. • must be of the same size

.

22. • ;

The

23. • the feet of the rods

should

24. • must be located on the same surface (

e.g.

25. • for example, floor)

.

26. • ;

In

27. • in complex scenes,

it is recommended to add

28. • we recommend adding more than three rods to increase the accuracy of the detection

.

29. • ;

In

30. • in the parts of the frame where the lens vertical distortion is observed, the rods

should

31. • must be parallel to the nearest vertical objects (

e.g. door,

32. • for example, door or wardrobe)

. Note

title	Attention!

For

33. • ; for portrait-oriented scenes (such as corridor, shopping aisle, warehouse aisle,

etc.

34. • and so on), the arrangement of the rods

should

35. • must form triangles.

      Info
      title Note

 For your

36. • For convenience of configuration, you can

click the Image Removed button and configure the mask on a still frame. To undo

37. • "freeze" the frame. Click the Image Added button. To cancel the action, click this button again.

Info
title Note
To delete the rod, click the Image Removed button.

38. In the Distance sensitivity field (1) enter a triggering distance value in meters between people. The detection tool will trigger if the distance between people in the frame becomes equal or less than the specified value. The value should be in the range [0; 20].
    Image Removed

39. In the Leveling rod length field (2) enter the average height of a person (in meters), according to which the leveling rods are set. The value should be in the range [0; 3].

40. Set the minimum number of frames in which the detection tool should detect close-standing people in order to trigger (3). The value should be in the range [1; 100].

41. • The area for detection is displayed by default. To hide it, click the Image Added button. To cancel the action, click this button again. To delete a rod, click the Image Added button.

      
      
42. Click the Apply button.

Configuration of the Close-standing people detector Configuring the frame perspective for the Close-standing people detection tool is complete.