Open topic with navigation
SeeTec
Analytics
The OnSSI server must send the camera images to the OnSSI Analytics module for intelligent video analysis. Based on the analysis, alarms can be triggered. The OnSSI solution offers different OnSSI
Analytics packages:
- OnSSI Analytics Basic
- OnSSI Analytics Premium
- OnSSI Analytics Enterprise
For the differences between the packages see Overview of the OnSSI Analytics packages.
To configure a OnSSI Analytics module, you first have to create it in the OnSSI
VA administration tool (see Ocularis Recorder VA Administration Tool)
Adding the OnSSI Analytics module
- Select Other
hardware and create a new object.
- Enter a Name
and select " OnSSI
Video Analytics" as Manufacturer.
- Select " OnSSI Analytics Basic",
" OnSSI Analytics
Premium" or " OnSSI
Analytics Enterprise" as Type.
- Select the Video
Analysis module to specify which video analysis channel is to analyze
the image data.
- Click OK.
General
- Activate the module
and, if necessary, alter the name.
- Select the video analysis
module to specify which video analysis channel is to analyze the image
data.
- Select the camera for
the intelligent video analysis.
Privacy masking must not be used on the selected
camera. OnSSI Analytics
devices cannot be copied. If the camera is rotated after configuration,
the Video Analysis module will terminate recognition in some circumstances.
- Select the Video
classification to specify which profile is to be used for image
transmission to the video analysis channel. This setting applies only
to multistreaming.
OnSSI recommends using an additional
stream from the camera for video analysis. A CIF resolution is generally
adequate for analysis (compression: 20%). Select larger resolutions only
after consultation.
- Specify the
Frame rate for analysis to set the number of images per second
to be sent to the video analysis channel.
OnSSI recommends a frame rate of 12
frames per second (fps).
- The Channel
state shows the current state of the OnSSI
Analytics channel. The following states are possible:
- Known scene (Good View): The OnSSI Analytics channel is running. The video data is being analyzed.
- No signal (Bad Signal): The OnSSI Analytics channel is running. The video data is not being analyzed. (Possible causes: the image quality is too poor, the contrast may be too low or the image is too dark.). An alarm scenario can be started in this state.
Select the Video analysis failure due to insufficient light or obstruction option as the trigger event for the alarm scenario. - Invalid scene (Invalid View): Unknown error. The video data is not being analyzed.
- Unknown scene (Unknown View): The current scene does not correspond with the acquired scene (e.g. the camera was rotated). The video data is not being analyzed. An alarm scenario can be started in this state. Select the "Video analysis failed due to incorrect camera position" option as the trigger event for the alarm scenario.
- Search for known scene (Searching for View): The OnSSI Analytics channel is in the acquisition phase. If the OnSSI Analytics channel is activated, the state changes to "Known scene", "Unknown scene" or "Unknown".
- Unknown: No connection to the OnSSI Analytics channel.
- Select the desired
Usage.
- Inside. All objects will be monitored inside of a room or buildingg
- Inside and people only. Only people will be monitored inside of a room or a building
- Outside. All objects will be monitored in open areas
- Outside, but no people. All objects except humans will be monitored in open areas
- Any. Any of the above applies
- Select whether the
analysis is to recognize people or vehicles.
- Click the Person
or Vehicle button to create an area in
the camera image to be used to calibrate the camera.
- For spatial location
of the vehicle click the Tires button and
define the position of the tires.
- In the specified area,
specify the position of the body parts (head/foot) or the vehicle's upper
point (e.g. roof).
Example
If a person is selected, place the head icon (red) at the highest point of
the body's image and the foot icon at the lowest point of the body.

- If a vehicle is selected,
place the cross icon at the highest point (roof) and select at least one
wheel.At least three persons or vehicles must be defined.
- To delete an area that
has been created, select it and click the Delete button.
- To delete all areas,
click the Delete all button.
- Activate Automatically
acquire new angles of view to store the new position of the camera
automatically as a new scene. An alarm
scenario can also be optionally started if the camera was rotated.
- Activate Force
new angles of view to prevent circumstances in which the video
data cannot be analyzed because the current scene does not correspond
to the acquired scene. Unlike the "Automatically
acquire new angles of view" option, you cannot start an alarm scenario
by activating the Force new angles of view option.
- Activate Export
single images in the event of an alarm to save the images on which
a motion or change is recognized separately as JPEG.
- Specify the time period
after which the exported images will be deleted. If
you enter "0" as the time period the images will not be deleted.
- Optionally enter the
generic parameters that are to be taken into account in the analysis of
the images. See Technical Guide or ask SeeTec support about
the use of generic parameters.
- Click the Add
new rule button and enter a unique name
for the rule.
- Select the rule type.
The following rules are available:
- Tripwire: As soon as the tripwire is exceeded in one or both directions, an alarm can be triggered.
- AOI (Area of interest): An area in the camera image that can trigger an alarm.
- Scene change: All changes in the visual field of the camera are detected if the camera image changes significantly.
A rule is not automatically activated after it
has been created. Activate the rule during configuration of the rule (see
below). Additional starting events can also be set for every rule.
Tripwire
- Activate the module
and, if necessary, alter the name.
- Activate Classification
of target to assign the recognized object types.
Multiple objects can be selected.
- Click the Tripwire
button to create a tripwire and record virtual tripwires in the camera
image.
- Click the MultiSegment
TW button to create multi-segment tripwires.
Multi-segment tripwires can be created at an angle.
- Draw the virtual tripwires
in the camera image.
Generate the corners of the tripwire by clicking on the camera image and
then drawing them.
- If you have created
a multi-segment tripwire, close it with a double-click.
- Select the parameters
for the directions from which the object is to be recognized.
- Select the filter
for the minimum size and the filter for
the maximum size to specify the size of the object that is to be
recognized.
- Activate the filters
with which the object is to be displayed.
The color highlighting makes it easier to recognize the sizes of target
objects in the foreground or background during configuration.
- Select the additional
filters:
- Salience: Moving objects ignored, e.g. reflections, falling leaves or water motions.
- Maximum size change (%): Fast-changing objects are filtered, e.g. shadows.
Example
A shadow
has a size of 100% in the first image, the size change is set to 50%.
If the shadow becomes larger, it is ignored until the size change exceeds
150%.
If the shadow becomes smaller, it is ignored until the size change falls
below 50%.
- Click the Delete
this rule button to delete the currently displayed rule.
- Click the Reset
counter button to reset the object counter.
This function is only available if the
SeeTec Counting
Suite is configured.
AOI (Area of interest)
- Activate the module
and, if necessary, alter the name.
- Activate Classification
of target to assign the recognized object types.
Multiple objects can be selected.
- Select the trigger
types to specify which object behavior is to trigger the event.
- Exiting: Object exits the analysis area (including partially) and is still visible
in the image.
- Entering: Object enters the analysis area (including partially) and was still visible
in the image beforehand.
- Disappearing: Object exits the analysis area and is then not visible in the image.
- Appearing: Object is in the analysis area and was not visible in the image beforehand.
- Click the AOI
button and specify the image detail.
- Close the image detail
with a double-click after it is marked out.
- Select additional
parameters for object recognition.
- Enter a time period for Time of loitering (s)and
Left behind since (s) after which a signal will be sent to trigger an alarm.
- Loitering: Object has been in the analysis area for a longer time (over x seconds).
- Left behind: Object is left behind in the analysis area for longer than x seconds (including partially).
- Select the plane to specify the plane on which the object is found.
This shows the differential background for the object recognition.
- Ground
plane: Recognition is to be on a horizontal area, e.g... a long corridor. The ground plane can be viewed as a carpet in the corridor that the object
is in contact with.
- Image
plane: Every motion is to be detected. A detection at image plane can best be compared with traditional motion detection. However, detection
at image plane also includes object classification.
- In the layer,
select whether detection is to be performed in an image
detail or in the full image.
Detection in the full image should be
used if all unspecific motions are to be detected (person enters or exits
the camera area). Because all motions are detected, it is also possible
to receive a number of unwanted alarms. If detection is to be performed
in the full image, some object behavior cannot be used and is therefore
grayed out.
- Select the filter
for the minimum size and the filter for
the maximum size to specify the size of the object that is to be
recognized.
- Activate the filters
with which the area is to be displayed. The color highlighting makes it easier to recognize the sizes of target
objects in the foreground or background during configuration.
- Select the additional
filters:
- Salience: Moving objects ignored (e.g. reflections, falling leaves or water motions).
- Maximum size change (%): Fast-changing objects are filtered, e.g. shadows.
Example
A shadow has a size of 100% in the first image, the size change is set to 50%. If the shadow becomes larger, it is ignored until the size change exceeds
150%. If the shadow becomes smaller, it is ignored until the size change falls
below 50%.
- Click Delete
this rule to delete the currently displayed rule.
- Click the Reset
counter button to reset the object counter.
This function is only available if the
SeeTec Counting
Suite is configured.
- Apply the set values if you want to make further
settings.
- Save
the set values to apply the values and conclude input.
Scene change
- Activate the module
and, if necessary, alter the name.
- Select the scene
change that is to trigger an event. All
changes in the visual field of the camera are detected if the camera image
changes significantly.
- Any change: Combination of available options of the selection.
- Lighting on: Change from dark to light
- Lighting off: Change from light to dark
- Lighting on or off: Change from light to dark or dark to light
- Unknown process: Unclassifiable change
- Camera movement: Camera is rotated
- Apply
the set values if you want to make further settings.
- Save
the set values to apply the values and conclude input.
Open topic with navigation