Pelco Analytics - How to configure analytic profiles and behaviors on Sarix cameras.

Issue

  • The configuration method or analytic settings are unknown.
  • Analytics are not functioning correctly.

Product Line

Pelco Cameras

Environment

  • Sarix Legacy
  • Sarix Enhanced
  • Sarix Professional

Cause

Incorrect configuration of analytics.

Resolution

 

 In order to enable analytics, you must create a profile.  Profiles define zones in which analytic behaviors are deployed. Depending on the camera model and settings, a maximum of three can be run as part of a single profile. The camera can store up to 25 profiles but can only run one at a single time.

Here are few general tips when configuring Sarix analytics:

  • Default settings, or settings that are pre configured by default, can help alleviate analytic configuration issues and is also useful to determine if target detections are being achieved.
  • Once an analytic response has been established, further tuning should then be performed to help reduce false positives and false alarms.
  • Adjust one parameter or setting at a time and test the analytic after a change has been made to determine if the adjustment has made any difference.
  • Avoid any unnecessary adjustments to the analytic settings. 
       


To create a profile:

  1. Log into the camera that will be set up for analytics.



     
  2. Access the Analytic Configuration in Events.

    Analytic Configuration Events Menu
  3. Create the new profile.


     
  4. Adjust each profile setting as required for the chosen behavior as followed below:

A) Profile Settings: Scene Type (lighting variable)


 

  • Indoor: Scenes with constant light.
  • Outdoor: Scenes with variable light.
     

 Explained: Analytics work best under conditions where lighting and contrast are consistent.
 

  • Behaviors that use Scene Type: Adaptive Motion, Loitering Detection, Object Counting, Camera Sabotage. 
  • Behaviors that do not use Scene Type: Abandoned Object, Stopped Vehicle, Directional Motion, Object Removal, Auto Tracking.  
     

B) Profile Settings: Background (movement in the background)

  • Still: little movement in the background.
  • Noisy: "Busy" scenes where there is movement generated by non-targeted objects.
     

Explained: Noisy setting should reduce false positives in busy scenes.
 

  • Behaviors that use Background: Adaptive Motion, Loitering Detection, Object Counting, Camera Sabotage.  
  • Behaviors that do not use Background: Abandoned Object, Stopped Vehicle, Direction Motion, Object Removal, Auto Tracking.
     

C) Profile Settings: Fine Tuning (speed at which movement is blended or assimilated into the background).


 

  • Aggressive: Refreshes or updates the background at a faster rate (slow to medium speed moving objects may become part of the background).
  • Conservative: Blends in moving objects into the background at a slower rate.
  • Normal: Default setting, middle range assimilation speed.

    Explained: Aggressive for differentiating fast targets from slow non-targets. Conservative when target movement is variable. Normal when movement of target speed is constant.
     
  • Behaviors that use Fine Tuning: Adaptive Motion, Loitering Detection, Object Counting, Camera Sabotage.

Behaviors that do not use Fine Tuning: Abandoned Object, Stopped Vehicle, Direction Motion, Object Removal, Auto Tracking.


D) Profile Settings: Sensitivity (Adjusts responsiveness to foreground events against the learned background scene).


 

  •  Scale adjustment from 1 to 10
     

Explained: Similar to Motion Detection, adjust the mid-range values initially then fine tune better accuracy.

To minimize false negatives, move the setting one to two steps higher
 

  • Behaviors that use profile sensitivity: Adaptive Motion, Loitering Detection, Object Counting, Camera Sabotage.
  • Behaviors that use individual zone sensitivity setting and ignore profile sensitivity: Abandoned Object, Stopped Vehicle, Directional Motion, Object Removal.
     

E) Profile Settings: Calibrate Scene(sets height and width of target objects in different areas of the scene).




Explained: Adjusts how the camera perceives depth perception, distance, and perspective.

The most critical aspect of configuring this setting is to use the same object for calibration and ensure that the object being used is similar in size to the actual target.

Example: The object being used for calibration is a person that is similar in size to the targeted individual(s) or object(s). Make sure the same person or object is used to calibrate the entire scene (near left, near right, and distance boxes). This will assist with the camera's two dimensional depth perception in relation to size and distance. 
 

  • Set Height: Input the person or object's height in the "Real World Height” field. 
  • Set Width: Use a consistent width, running from near to far in the camera's field of view. Set the far width first, then the near width. Be sure to enter both height and width when calibrating the scene.
     
  • Behaviors that use Calibrate Scene: Adapter Motion, Loitering Detection, Object Counting, Camera Sabotage, Abandoned Object, Stopped Vehicle.
  • Behaviors that do not require calibration: Directional Motion, Object Removal, Auto Tracking.


F) Select Behaviors

  1. Under Select Behaviors,  highlight or click the predefined behavior(s) from the list. Settings for the selected behavior will be enabled. You can activate up to three behaviors per profile, provided that the camera has sufficient memory.
  2. Check the Activate Behavior box below the settings for the selected behavior.


     
  3. To enable the analytic profile, click Run under the newly created profile. There will be an arrow placed next to the profile when the profile is active and running.