Abstract
Augmented reality (AR) is a hot field now emerging, which has the characteristics of virtual–real fusion display. It is often used in the control field, the main technical difficulties of which are the low precision of outdoor tracking registration and the low real-time quality. So, this paper presents a target tracking method, which combines the particle swarm algorithm and the Kalman filter. The method uses Kalman filter to forecast the position where the center of a target locates in the image, thus greatly reduced the search scope, and the position is used to set target search area as the center. And then, matching target templates and candidate area with the characteristics of target ensures the instantaneity and accuracy of registration and tracking.