Newton T-DAR T1010MT, Инструкция по установке и эксплуатации

Страница: 76 / 112

69
***IMPORTANT NOTE***
Calibrations must be performed when the scene is completely stable. The
scene must be stable for several seconds after the button is clicked, until the
image stabilizes at a black level.
Sensitivity
The sensitivity section contains slider bars for maximum head size, cart sensitivity,
and crawler sensitivity. Adjacent to each of these sliders is a numeric box that shows
the numeric equivalent for the slide position. The values may be adjusted by either
manipulating the slider or through direct entry of the values in the numeric box. The
slider bars are detailed as follows:
About Settings:
The use of stereo machine vision analysis allows the system to assign a fairly
constant numeric value to an object regardless of how far or close the object is to
the camera lens. This number represents the amount of pixels the system needs to
see to place a
cross
-hair on an object
Minimum Head Size
The minimum head size field (above L2, L3, and L4) controls the value (in pixel area)
that determines the smallest object that can be called a head without incorrectly
identifying carried objects as a head. To stop a small object from being detected as
the head of a person, set the minimum head size (in pixel area) to be larger than this
small object. Setting this value too small may cause the system to perceive a carried
object as the head of a person. Setting this value too large will cause the system to
ignore persons who’s heads do not meet this minimum requirement. If the value for
minimum head size is set to low, small objects can be incorrectly considered the
head of a person. A standard value is 350.
The monitor output can be configured to show the number of persons detected
under a camera head by displaying a “Head Count” value and showing a cross over
every object that is considered the head of a person. Viewing this output is
accomplished by using the ‘Show Camera Views’ option from the monitor tab, which
also provides a tracking image. The public door or the secure door may be
individually selected for monitoring.
L2, L3, and L4
The values L2, L3, and L4 represent various horizontal
cross sections of a person’s
body at different heights, measured from the top of the head. The value L2
corresponds to the area (in pixels) one inch down from the top of a persons head.
The value L3 corresponds to the area two inches down from the top of a persons
head. The final value L4 corresponds to the area nine inches down from the top of a
persons head (shoulder area roughly). Typical values for these
fields are 2500, 3200,
and 3800 (units are in pixels).