There are two kinds of domain of shot
segmentation: the non-compressed domain and the
compressed domain. This paper discusses the shot
segmentation and the key frame extraction in noncompressed
domain. The method of shot segmentation
in non-compressed domain mainly includes pixel and
block level comparisons, and histogram comparison
methods [4]. Unfortunately, these traditional methods
of shot segmentation are either very sensitive to
movement of the object and the camera, or demanding
too much calculation. So they are not suitable to be
applied to lane video surveillance system.