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 the block matching motion estimation of the video frame image,the movement direction and the position of marker were predicted through the moving correlation of consecutive frames.So the method reduces the time complexity while the augmented reality system detects the marker with full frame.The method on the basis of ARToolKit can significantly improve the time performance of tracking the marker.
出处
《河南科技大学学报(自然科学版)》
CAS
北大核心
2013年第2期57-60,3+2,共4页
Journal of Henan University of Science And Technology:Natural Science
基金
国家科技支撑计划子课题基金项目(2011BAH02B04)
北京市自然科学基金项目(4112021)
北京联合大学"启明星"大学生科技创新项目(12205994104)
关键词
增强现实
块匹配
跟踪注册
跟踪预测
augmented reality
block matching
tracking registration
tracking prediction