Documentation for Axxon One 2.0 version. Documentation for Axxon One 3.0 version is also available.


You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 2 Next »

Hardware requirements for neural analytics operation

Requirements for the operation of Object Tracker and its sub-tools are given in the table:

Camera
  • Resolution must be at least 640x360 pixels. We also don't recommend using a resolution higher than 800×600

    Attention!

    Higher resolution doesn't increase the detection quality, but significantly increases the CPU and RAM load.

  • FPS is at least 6
  • Tracker-based analytics works with both colorless (gray) and color images. If you use neurofilter, we recommend using color cameras. When you use black-and-white cameras, the recognition quality can be noticeably worse
  • Camera shake shouldn't result in image offsets greater than 1% of the frame size
Lighting
  • Moderate lighting. In conditions of insufficient or excessive lighting (night or light-striking), stable operation of the analytics isn't guaranteed
  • No abrupt changes in lighting
Scene and camera angle
  • Moving objects are visually separable from each other
  • The background is mostly static and doesn't change abruptly
  • Moving objects are minimally obscured by static objects in the scene (columns, trees, and so on)

  • Analytics may not work correctly if:

    • the analyzed scene have reflective surfaces and sharp shadows from moving objects;

    • there are extended monochromatic objects in the frame

  • When using the Abandoned Object detection tool, abandoned objects may be obscured by moving objects no more than 10% of the time

Object images

  • Image isn't noisy and isn't distorted by artifacts of the compression algorithm

  • The width or height of objects in the image should not:

    • be less than 1% of the frame size (at resolutions no less than 1920x1080 pixels) or 15 pixels (at lower resolutions)

    • exceed 75% of the frame size

  • When you use a neurofilter, the minimum width or height of an object must be at least 5% of the width/height of the frame

  • When using the Abandoned Object detection tool:

    • abandoned objects must be clearly visible in the image

    • the width and height of objects in the image should not exceed 25% of the frame size

  • The speed of movement of objects in the image should not be less than 1 pixel per second

  • To the object to be detected, the duration of its visibility must be at least 8 frames

  • When the object moves between two adjacent frames in a certain direction, its distance of movement should not exceed its size. This condition is necessary for correct calculation of the object’s trajectory (track).

Provided that the specified requirements are met, the Abandoned Object detection tool is guaranteed to:

  • detect 92 abandoned objects out of 100;
  • keep false positives to 20 out of 100.
  • No labels
Error formatting macro: children: java.lang.NullPointerException