Documentation for Axxon One 2.0. Documentation for other versions of Axxon One is available too.

Previous page Image requirements for the Object tracker and its sub-detectors  Configuring the Object tracker Next page

Configuring the Object tracker

To avoid false positives from the sub-detectors of the Object tracker, the following recommendations must be met:

  1. We recommend configuring the parameters of the Object tracker empirically, checking the quality of its operation at each step.

  2. Parameters that strongly influence the quality of operation:
    1. Max. object height, Max. object width, Min. object height, Min. object width (see Configuring the minimum and maximum size of detected objects):
      1. The values of the Max. object height and Max. object width parameters must be selected so that they are slightly lower than the size of a typical object in the image.

      2. The values of the Min. object height and Min. object width parameters must be selected so that they are slightly larger than the size of a typical object in the image, taking into account its shadow.

    2. Motion detection sensitivity, Abandoned object detection sensitivity and Auto sensitivity:

      1. We recommend using the Auto sensitivity parameter that ensures the quality of the Object tracker performance in highly variable lighting conditions.

      2. If lighting is stable, we recommend selecting the values of the Motion detection sensitivity and Abandoned object detection sensitivity parameters empirically. For detection of objects with low contrast, the recommended sensitivity value is 35, for contrast objects—15.

    3. If you decrease the value of the Time of object in DB parameter, generation of events on false static objects will be excluded. If you increase the value, in some cases the object track will be saved even when it temporarily disappears or overlaps with other objects.

If the requirements for the Object tracker are met (see Video stream and scene requirements for the Object tracker and its sub-detectors, Image requirements for the Object tracker and its sub-detectors), the following performance parameters of the Abandoned object detector are guaranteed:

  • out of 100 abandoned objects, the detector detects 92 objects;

  • 80 detection events out of 100 will be caused by real abandoned objects, and 20 detection events will be caused by other changes in the monitored scene.

  • No labels