To help the analytics detect people at varying distances, you can configure the
perspective of the scene, including the height of near and far objects.
What you should know
Use the white grid as a visual reference for how the perspective of the scene is
interpreted. This grid changes according to the size and placement of the near and far
objects, and should match the ground plane.
NOTE: Depending on which scenario type you are configuring, some
settings might be unavailable or hidden if Advanced mode
is
disabled.
Procedure
-
From the Config
Tool home page,
open the Video task.
-
From the area view, select a camera that you applied an analytics scenario
to.
-
Click , and then click the scenario you want to configure.
-
Click the Perspective tab.
-
Use one person in the video as a size reference to configure the near and far
objects:
-
When the person is near to the camera, click and drag the
Near object over the person in the
video.
-
Adjust the Near object to match the height of
the person by clicking and dragging the white dot on the object.
-
Repeat the previous substeps using the Far
object when the same person is far from the
camera.
-
In the Perspective section, configure the following
settings as required:
NOTE: Changing the block size affects the size of the blocks displayed when you
enable the Show movement grid option in the area
configuration.
- Block size
- The size of analysis units relative to the specified object or
person size in the image. If you set a low value, the required computing power
increases, but the distinction between different object sizes in the image is more
accurate.
- Min. block size
- The minimum size of analysis units relative to the height or
width of the video, depending on which measurement is shorter.
- Object height
- The real-life height of near and far objects that you want to
monitor. The default setting is 1.75 m (5.7 ft), the average height of a person.
Tip: Ensure that the Object height defined in the
Perspective section matches the height represented by the near
and far objects in the video.
- Scene depth
- The scene depth distance in the view direction of the camera. The
scale represents the size of an average person.
- Low block resolution
- When this option is enabled, the accuracy of motion
detection per block is reduced. The required computing power is also reduced.
-
Click Apply.