Configuring the line for the People counting scenario - KiwiVision™ 4.6.1 | Security Center

KiwiVision™ User Guide for Security Center

Content type
Guides > User guides
Last updated

You must define the line that must be crossed so that people can be detected entering and exiting an area.

What you should know

  • You can visualize the people counting statistics in Security Desk dashboard widgets.
  • When creating event-to-actions for cameras that trigger the People counting event, you can specify In or Out as conditions if you want to trigger actions based on direction.


  1. From the Config Tool home page, open the Video task.
  2. From the area view, select a camera that you applied an analytics scenario to.
  3. Click Video analytics > People counting, and then click the scenario you want to configure.
  4. Click the Line tab.
  5. Define the line as required:
    1. Click and drag the white dots to adjust the length and position of the line.
    2. (Optional) Add or remove points:
      • To add a point, click on the line.
      • To remove a point, select it and click Remove selection ().
        NOTE: When a point is selected, it is yellow.
  6. In the Detection section, configure the following settings:
    Each scenario describes a different camera setup and context for analysis. Select the option that best describes your situation:
    The camera is at an angle, and the full view of people can be seen.
    Tilted (crowded)
    The camera is at an angle, there are many people in the scene, and they obstruct the full view of each other so that only their heads are seen fully.
    Top down (fast)
    The camera is facing directly down at the ground. The speed of the algorithm is prioritized and less CPU is used, but detections might be less accurate.
    Top down (accurate)
    The camera is facing directly down at the ground. The accuracy of the algorithm is prioritized, but more CPU is used.
    The accuracy threshold of the algorithm required for a person to be considered a detection in a video frame. Detections are displayed as colored circles; each color indicates a different detected person.
  7. In the Direction section, select the direction that triggers a people counting event.
  8. Click Apply.