A Traversing and Merging Algorithm of Blobs in Moving Object Detection
Yong Zhou, Yi Li · Applied Mathematics & Information Sciences · 2013
In the frame of surveillance video, the colour of both moving objects and background distribute randomly. Consequently, the foreground image obtained by background subtraction or temporal differencing always contains many discrete blobs, which are isolated but actually belong to the same object. These blobs influence the effi ciency of later process such as recognition, classification, and activity analysis. In order to organize the foreground blobs, a trav ersing and merging algorithm of blobs is proposed in this paper. The experimental results show that the proposed algorithm can merge blob effectively.