The correlation-matching algorithm determines the matching degree of model and grabbed image by calculating their correlation value. Because the algorithm has high precision and strong adaptability

and is robust to the linear change of grayscale

it is widely used in projects as a classic matching algorithm. But correlation-match computing is very time-consuming and difficult to satisfy the real time request. Besides

when the target is only small part in the model

it is hard to confirm the model position. In this paper

an adaptive model-based real time tracking algorithm is introduced

in which the grabbed image is segmented by threshold and the noise is removed. Then a contour segment method is used to determine the size of model. The matching algorithm chosen for calculating the degree of match is a normalized correlation algorithm accelerated by pyramid algorithm. The results show that when the target is very small in the model

the tracking algorithm has better matching precision and real time.