Video Analytics best practices

Camera placement

Camera location

Video Analytics detects objects and classifies them as people, animals, and vehicles. Position the camera with a view that can see important activity from people, animals, and/or vehicles.

Lighting

Be aware of any bright lights (i.e., motion-triggered floodlights) near the camera that may attract insects or affect exposure and IR settings. Mount the camera in such a way that a vehicle’s headlights do not shine directly into the camera.

Avoid reflective surfaces (i.e. overhangs/eaves, walls, etc) that will reflect IR lights at night and make it difficult to see objects further away as reflective surfaces become overexposed.

When to use a Ground Zone rule

Use a Ground Zone rule when the object to be recorded lingers in a specific place for too long.

Verify that the Ground Zone is fitted to the ground or floor where the object's feet will be.  For more information about configuring a Ground Zone rule, see Configure rule - Ground Zone.

When to use a Tripwire rule

Use a Tripwire rule when the object to be recorded passes through a specific place.

Verify that the Tripwire is placed where the object's feet or wheels will cross the threshold. For more information about configuring a Tripwire rule, see Configure rule - Tripwire.

To ensure rule accuracy

• Avoid drawing Tripwires or Ground Zones over objects that might move by themselves (e.g., trees, bushes, or other objects that move in the wind).
• Only draw the rule where the feet or wheels of an object should cross.
  • Do not increase the size of the loiter zone or tripwire to cover the whole of the target.
  • Do not include furniture, walls, trees, or other structures in the rule/selection.
     

    Bad ground zone example	Good ground zone example
    Including the bush and garage door in the ground zone can make it difficult for the Video Analytics to positively classify subjects. In this case, reduce the size of the ground zone to only cover the driveway.	This is a properly drawn ground zone that only covers the driveway, and excludes the bush and garage door.

    
     
• Create up to three Video Analytics rules to capture separate areas of interest (e.g., a camera that monitors two entrances might need two tripwire rules, each one covering one of the entrances).
• After observing what the rule detects over a few days, consider adjusting either the Minimum Object Detection Size slider or recalibrating the camera.
   

  Note: Certain video devices have a calibration process that replaces the Minimum Object Detection Size setting. Minimum Object Detection Size will not appear on this page for those cameras. For more information about the calibration process, see Calibrate a device for Video Analytics.

   
  • When adjusting the Minimum Object Detection Size slider:
    • Increasing the minimum object size can reduce false detection.
    • Decreasing the minimum object size can detect smaller targets.