CHeKT Pro AI Camera Series - Analytics Manual
The latest licenced plug-in is a set of analytical tools loaded on our cameras that provides the means to perform advanced
analytics, reduce false alerts and customize when events occur. This article will describe how to enable and configure the features of our video analytics to detect events of interest and trigger actions to react to those events.
Configuration
By default, the VCAedge plug-in is disabled. Activate
it to enable the plug-in.
General Setting
• Enable: Defines the current state of the plug-in features. Select On and Apply to turn the features on. Select Off and Apply to turn the features off.
Tracker Engine
Whether the tracker engine setting is enabled depends on the
hardware platform and active license.
• Deep Learning Object Tracker (DLOT) is designed for accurate detection and tracking of people, vehicles, and other key objects in challenging environments where motion-based tracking methods would struggle. It is based on a classification and detection model, providing the location of an object in the field of view
• Deep Learning People Tracker (DLPT): is designed to track people in dense and crowded environments where potentially only the head of a person is visible. Highly accurate and optimized for people counting and behavior detection.
Note: The menu system will reload after each apply to reflect the features that are available, the features that are available will depend on the license that has been applied.
*Warning: If you change the tracker engine, all settings will be reset to the factory default settings.
Rules
Rules are used to react to events within a scene and
trigger actions. To manage the rules, navigate to the rules tab from the CHeKT
AI menu. The rules page displays a live view from the camera and a table of the rules that have been defined for the camera. Add can be used to add additional rules, Modify is used to change the settings on a selected rule and Delete will remove the selected rule.
The types of rules available include:
• Area Detection – This rule triggers an event when an object is first detected in a particular zone.
• Line Crossing – This rule triggers an event when an object is first detected crossing a particular line.
• Object Enter – This rule triggers an event when an object crosses from outside a zone to inside a zone.
• Object Exit – This rule triggers an event when an object crosses from inside a zone to outside a zone.
• Object Appear - This rule triggers an event when an object starts to be tracked from within a zone.
• Object Dwell – This rule triggers an event when an object is present in a particular zone for a pre-defined period.
• Direction - The direction rule triggers an event when an object crosses the detection line in a particular direction and within the acceptance parameters.
• Tailgating - The tailgating rule triggers an event when objects cross over a line within quick succession of each other.
• Non-Detect Zone - Non-detect zones can be used to exclude areas of the scene from being analyzed. This can be used to reduce false triggers that can be caused by moving foliage or busy scenes.
Note: Rules cannot be deleted if they are linked to an action or counter. Remove these links before attempting to delete.
Show Burnt-In Annotation: Use this option to show analytic data on the camera
view, select Start to show the data
and Stop to hide the data
Note: The burnt-in annotation feature needs to be enabled for this option to function, this does not effect the processing of analytics but the annotation requires more resource from the camera and is not on by default.
Adding Rules
- Select Add to display the available rules.
- Select the rule to be added.
- Configure the rule accordingly.
- Select Save to save the rule.
Example | Area Detection: This rule will create a zone and overlay it on the live
view; the zone can be reshaped accordingly. Selecting a grey node will split
the segment and create a more complex shape, to remove a segment select the
minus sign next to a red node:
Note: The rule will trigger in the same circumstances as the Enter and Appear rule, the choice of which rule is most appropriate will depend on the scenario.
Rule Property
• Name: Defines the name of the rule.
• Object Filter: Allows the rule to be configured to only trigger based on an objects classification (e.g. person, car), any combination of the available options is possible.
Note: The available classifiers are different depending on the tracker engine.
The Event Actions configuration are optional and not required for the transmission of analytic data from the camera to the CHeKT Bridge.
Event Actions
• Event Notifications: Define the action that will occur on an event being triggered.
– TCP Event: Enables/disables the triggering of the TCP notification when an event occurs.– HTTP Event: Enables/disables the triggering of the HTTP notification when an event occurs.
The following applies for the notification methods selected.
• Triggered By: Define when the notification will be sent.
– Object: Send notification for each object triggering the rule. Choose between the begin of the object triggering the rule, as it enters the zones. and/or the end of the object triggering the rule, as it leaves the zone. A notification will be sent for each object triggering the rule.– Rule: Send a notification every time the rule is triggered, from the begin point of the first object to trigger the rule to the end point of the last object to trigger the rule. A notification will be sent for each triggering of the rule, regardless of the number of objects triggering at any one time.
Note: Remember to configure the TCP/HTTP notification actions for the action rules feature to function. Any combination of the available options is possible.
• Action Rule: Allows actions defined in the camera, to be triggered when an event occurs.
• Convert VCA to MD: Allows the event to be passed to the cameras event engine which is compatible with ONVIF.
Save or Cancel
Click Save to save the current settings or Cancel to
return to the rules screen without saving any changes
Counters
Counters can be configured to count the number of times
a rule is triggered, for example: the number of people crossing a line. To manage the counters, navigate to the Counters tab from the CHeKT AI menu. The counters page displays a live view from the camera and a table of the counters that have been defined
for the camera, Add can be used to
add a counter, Modify is used to
change the settings on a selected counter and Delete will remove the selected counter.
A counter is designed to be utilized in the following way:
• Increment: Whereby the counter is incremented on each trigger of an event.
• Decrement: Whereby the counter is decremented on each trigger of an event.
• Occupancy: Whereby the counter reflects the number of objects that are currently triggering the attached rule(s). It is available for a rule that links to a polygon zone.
More than one counting line rule can be assigned to a counter input. This allows, for example, the occupancy of two counting lines to be reflected in a single counter or more than one entrance / exit gate to be assigned to a counter.
Counters should not be used for occupancy and increment / decrement at the same time.Show Burnt-In Annotation: Use this option to show analytic data on the camera view, select Start to show the data and Stop to hide the data
Note: The burnt-in annotation feature needs to be enabled for this option to function; this does not effect the processing of analytics but the annotation requires more resource from the camera and is not on by default.
Adding Counters
- Click Add.
- Position the counter on the screen; this is just for reference and does not effect the counter feature.
- Define a name for the counter.
- Select Add Input against the desired action (increment or decrement) and select the desired rule that the counter will link to.
- When the selected rule triggers an event the counter will change its value accordingly.
- Click Save to save the counter.
Example:
Counter Property
• Name: Defines the name of the counter.
• Increment: The counting line rule which, when triggered, will add to the counter.
• Decrement: The counting line rule which, when triggered, will subtract from the counter.
• Occupancy: Sets counter to current number of the rule’s active triggers.
Select Add input to show a list of available counting lines that can be added. Select the cross next to rules already added to remove them from the counter.
• Counting Value: Defines to counter threshold that will cause an event to be triggered.
Event Actions
• Event Type: Define the action that will occur on an event being triggered.
• Action Rule: Allows actions, defined in the camera, to be triggered when an event occurs.
– TCP Event: Enables/disables the triggering of the TCP notification when an event occurs.– HTTP Event: Enables/disables the triggering of the HTTP notification when an event occurs.
Note: Remember to configure the TCP/HTTP notification actions for the action rules feature to function. Any combination of the available options is possible.
• Convert VCA to MD: Allows the event to be passed to the cameras event engine which is compatible with ONVIF.
Save or Cancel
Click Save to save the current settings or Cancel to return to the rules screen without saving any changes
Burnt-in Annotation
The Burnt-in Annotation feature allows analytical data to be burnt in to the raw video stream of the camera. Annotations can include tracked objects, counters and system messages.
Note: - To display object parameters such as speed, height, area and classifications, the device must be calibrated. - The stream intended for use with burnt-in annotation must have a resolution lower than 1920x1080, this is a limitation of the camera hardware. - In order for the show annotation function of Rules and Counters to work, the burnt-in annotation feature must be on and configured.
Burnt-in Annotation Setup
- Turn on the feature by select ON next to Activation.
- Select the stream/channel that will display the burnt-in annotation.
- Select the metadata that will appear on the stream.• Display zones: Displays the zone outline on the video feed.• Display objects: Displays the object bounding boxes on the video feed.– Object ID: Displays the unique ID of each tracked objects.– Object Confidence: Displays the object classification and its confidence value.– Object Dwell Time: Displays the elapsed time since the object started being tracked.– Object Color: Displays the top four colors that which make up more than 5% of the object are represented.– Internal State: visualizes additional tracker annotations.– Only Alarmed Objects: Displays only bounding boxes of objects which have triggered a rule.• Display event message: Displays event messages on the video feed.• Display system message: Displays the system messages on the video feed.• Display line counter: Displays the line counter information on the video feed.• Display counter: Displays the counters on the video feed.
- Click Apply to save.
Note: The calibration process must be completed for information such as speed, height, area and classification to appear.
Tamper
The Tamper feature is intended to detect camera tampering events such as bagging, defocusing and moving the camera. This is achieved by detecting large persistent changes in the image
Tamper Setup
• Enable: Defines the current state of the plug-in features.
– On: Enables the tamper feature.– Off Disables the tamper feature.
• Duration: Defines the the length of time, in seconds, that the image must be persistently changed before the alarm is triggered.
• Area Threshold: Defines the percentage area of the image which must be changed for tampering to be triggered.
• Suppress Alarm on light: In scenarios where sudden changes in lighting such as switching on/off indoor lighting can cause false tamper events.
– On: Enables the suppressing feature.– Off Disables the suppressing feature.
Note: The option will reduce sensitivity to genuine alarms and should be used with caution. Remember to Apply changes for them to take effect
Event Action
• Enable TCP Event: Defines if a TCP notification will be sent when the tamper detection triggers.
– On: Enables the TCP event feature.– Off Disables the TCP event feature.
• Enable HTTP Event: Defines if a HTTP notification will be sent when the tamper detection triggers.
– On: Enables the HTTP event feature.– Off Disables the HTTP event feature.
• Action Rule: Allows actions, defined in the camera, to be triggered when an event occurs.
Click Apply to save the current settings.
Updates: Firmware & Plugin
For a copy of the full CHeKT AI Analytics manual please download the pdf file below
Related Articles
CHeKT 4MP 12vDC AI Turret Camera
Crafted with durability in mind, the 4MP AI Turret camera boasts a robust construction, ensuring its ability to withstand various environmental conditions. When used the CHeKT Video Bridge, this camera takes advantage of modern IP technology to ...Analytic Triggers: Using Camera Analytics
Analytic Events Supported analytic event triggers from a camera mounted to the bridge can generate alarm signals and video events. These events are based on a supported API connection with the camera. The network event message has topic information ...CHeKT Cloud AI Filtering: A Practical Setup Guide
CHeKT Cloud AI filtering is available to all CHeKT Dealers on the CHeKT Professional and Premium plans; it offers an additional level of false alarm prevention to be used in tandem with the intelligent analytic detection on the mounted camera. How it ...OPTEX InSight Cameras
OPTEX InSight Cameras The Optex InSight series are the new members of the OPTEX sensor family of outdoor detection. By combining the dual-tech sensors of an OPTEX sensor and a full HD 1080 camera, you now have a powerful Visual Verification PIR ...Integrating Cawamo with CHeKT
The Cawamo AI Server integrated with the CHeKT platform creates video AI event notifcaitons for alarm monitoring centers and customers using the CHeKT Mobile app. This integation allows sites to leverage Cawamo's advances AI as detection zones for ...