From Wikibooks, open books for an open world
Jump to navigation Jump to search

Description[edit | edit source]

This step intends to separate the pixels that belongs to tracked objects. The pixels kept are in white and the black pixels will be considered as the background. Most of the time, this step follows a preprocessing method like background subtraction.

Input[edit | edit source]

SwisTrack BackgroundSubtractionColor.jpg

A color image.

Output[edit | edit source]

SwisTrack Threshold.png

A binary (black and white) image.

Parameters[edit | edit source]

Threshold[edit | edit source]

A threshold value (between 0 and 256). If the tracked objects don't appear in the image, try to reduce the value. Otherwise, if parts of the background are selected, the value need to be increased.

Using average value of the channels[edit | edit source]

Whether to threshold the free channels independently with the same threshold value and apply an OR function between the resulting binary masks, or to compute the average of the three channels and then apply the threshold.

Invert threshold[edit | edit source]

Whether the threshold must be inverted or not.