Video object extraction and its tracking using background subtraction in complex environments

Background subtraction is an efficient way to localize and obtain the centroid of the connected pixels moving on the foreground despite the prior information of the scene. It is suitable under fixed camera arrangement, which incorporates many vision applications such as object tracking, human monitoring, etc. However, the moving object extraction task becomes sophisticated and challenging due to some annoying factors such as local motion in background (waving tree, rippling water, etc.), camouflage region, sleeping object, which in turn degrades the tracking performance. In order to alleviate these problems, an efficient background subtraction algorithm is proposed to support the object-tracking task under static and dynamic background conditions. The work is focus to realize the relevant moving blobs on foreground by aiding the proper initialization and updating of the background module in order to improve the tracking accuracy. It generates an initial motion field using spatial-temporal filtering on the consecutive video frames. The block-wise entropy is evaluated above a certain range of the pixels of the difference image in order to extract the relevant moving pixels from the initial motion field. A suitable threshold value is estimated to assign an appropriate label to the moving blobs on the foreground mask. Finally, an adapting Kalman filter is integrated to the object extraction module in order to track the object on the foreground. Extensive quantitative experiments prove that the proposed method competently handles the object extraction, which in turn improves the tracking task under static and dynamic background conditions