Byoung Muk Min , Hae Seok Oh

The KIPS Transactions:PartB , Vol. 13, No. 4, pp. 463-470, Aug. 2006
10.3745/KIPSTB.2006.13.4.463,   PDF Download:

Abstract

It is difficult to track and extract the movement of an object through a camera exactly because of noises and changes of the light. The fast searching algorithm is necessary to extract the object and to track the movement for realtime image. In this paper, we propose the correct and fast algorithm using the variable searching area and the background image change method to robustic for the change of background image. In case the threshold value is smaller than reference value on an experimental basis, change the background image. When it is bigger, we decide it is the point of the time of the object input and then extract boundary point of it through the pixel check. The extracted boundary points detect precise movement of the object by creating area block of it and searching block that maintaining distance. The designed and embodied system shows more than 95% accuracy in the experimental results.

Statistics
Cite this article