Automatic estimation of the brightness changes for background suppression methods used for video tracking of vehicles

One of the typical distortions in the background estimation methods is a change of lighting conditions, since each such change influences on the luminance of pixels in the captured images, which may be classified as the background. The global changes are relatively easy to compensate, but in practical applications the character of most of such changes is rather local. These changes may be caused e.g. by clouds, moving large objects, street lamps etc. Nevertheless, their influence on the results of the background estimation should be reduced therefore a local adaptive correction algorithm, applied as the pre-processing step, is proposed in the paper, assuming known geometrical configuration of the observed road