A real time change detection technique is proposed in order to detect the moving objects in a real image sequence. The described method is independent of the illumination of the analyzed scene. It is based on a comparison of corresponding pixels that belong to different frames and combines time and space analysis, which augments the algorithm’s precision and accuracy. The efficiency of the described technique is illustrated on a real world interior video sequence recorded under significant illumination changes.
Top2. Methodology
Image sequence consisting of N video frames is observed. The sliding mask is applied on every frame.
Skifstad and Jain, (Skifstad, 1989), use the ratio of pixel intensities in mask between two a reference and a current frame to estimate the pixel variance as follows:
,
.
(1)Here, pixel intensities within mask are denoted with for a reference, background frame that does not contain changing regions and with for a current frame where moving objects are being identified. The mean of the pixel intensity ratio within is denoted with. If (where is a suitable threshold), the center of the mask is marked as changing region.
Experiments in (Inigo, 1989), (Mecocci, 1989), (Rourke, 1990), (Fathy, 1995), (Corall, 1991), (Foresti, 1998), (Foresti, 1994) have shown that for significant illumination changes this method fails, i.e. some pixels are falsely assigned to changing regions. A modified technique based on adaptive coefficient for illumination compensation is proposed in this chapter. Pixel variance is estimated as:
(2)