Table of Contents
Advertisement
6.3 Detail-Setup
6.3.6 Expert:
The settings of the expert menu are pre-configured for the most typical applications. In
order to adapt the sensor in an optimal way to the real szenary the meaning of the
following parameters are descripted:
The setting of middle / long in connection with changing illuminations can cause
unmeant activations particularly in outdoor applications. Therefore this parameter has to
be set very carefully.
Binaryfilter: Defines the suppression of noise- and fail-influences against to the de-
tecting objects. The function can be set in four steps.
The size of proximate, detected cells grows by increasing this parameter.
In case of great detecting objects at fore- and background (!) the increment of the bi-
naryfilter value is highly recommend!
Trace: The sensor is analysing all current, detected objects regarding to their further
course. The length of the resulted motion vector is a measure for the distance covered.
(Trace). The value of the trace parameter defines, whether a recognized object causes
immediately an alarm (left position, trace is disabled), or after achievement of a mini-
mum trace distance (short, middle, long).
Objects, which are moving predominantly in one direction, can be unambiguously identi-
fied. On the other hand, objects that do not move in one direction such as a tree in the
wind cause only a local movement without a significant trace.
Note:
Increasing of the value trace extends the analysing time of the objects and the time until
alarm triggering!
- 31 -
Ringbuffer:
The DMD-4 scans the entire video
picture 8 times per second. Thus,
very fast picture changes can be
correctly detected.
If tiny objects move very slowly,
sensors with a measurement of a
fixed, short time interval may not
always trigger correctly between
two video pictures which are to be
compared.
The change in brightness as com-
pared to the time interval is not
large enough to exceed the trigger
sensitivity (see also "sensitivity",
page 23).
The DMD-4 solves this problem by
an extension of the time interval by
means of an internal circular buffer.
Now slow movements in the picture
can be correctly detected.
Advertisement
Table of Contents
