A new target tracking and positioning method for grayscale images
Kezhi Li, Shuang Cong, Yakun Pu · 2011
To solve the problem of target tracking and positioning that easily affected by illumination and shadows, we apply the 2-D gradient code matching technique to the problem of target positioning, to make use of the features that the gradient direction angles are relatively stable to illumination and shadows variation. At the same time, Kalman filter is exploited to predict the target positions, by which increase the efficiency of the large quantity of computation in global search. Meanwhile, the gradient code matching is implemented in a much smaller range around the predicted position, so that it could reduce the computational complexity significantly.