增强现实中跟踪预测方法的研究

摘要

为提高增强现实系统中标记跟踪注册的实时性,基于视频帧图像块匹配运动估计方法,通过连续帧间运动相关性预测标记运动方向,在预测方向区域检测标记,降低了全帧检测标记的时间复杂性.基于ARToolKit,验证了本方法可显著提高标记跟踪的时间性能.%For improving the real time of marker tracking registration in the augmented reality, based on thernblock matching motion estimation of the video frame image, the movement direction and the position of markerrnwere predicted through the moving correlation of consecutive frames. So the method reduces the time complexityrnwhile the augmented reality system detects the marker with full frame. The method on the basis of ARToolKitrncan significantly improve the time performance of tracking the marker.