An improved Kalman filter algorithm base on quaternion correlation in object tracking

In this paper, we proposed an improved Kalman filter algorithm based on quaternion correlation of color image. Kalman filter is used to predict the object motion direction and define the searching window. In the searching window, the correlation of object and the window image is calculated to get the accurate position of the object. When the object is covered partly, we can get good result, too. Experiments results showed that the accuracy of our algorithm is higher than mean-shift algorithm.