Study on registration method based on Gauss-Newton iteration algorithm for augmented reality system

被引:2
|
作者
Li, Y [1 ]
Liu, Y [1 ]
Wang, Y [1 ]
机构
[1] Beijing Inst Technol, Dept Optoelect Engn, Beijing 100081, Peoples R China
关键词
augmented reality; image grabbing; registration; Gauss-Newton iteration; binocular stereovision; computer vision;
D O I
10.1117/12.561131
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Three key technologies influence the performance of current AR (Augmented Reality) system, namely image grabbing, accurate registration and binocular stereovision. This paper studies a vision-based AR system and its setup, presents an image grabbing solution using an IEEE 1394 interface, discusses the binocular stereovision technology and develops an effective Gauss-Newton iteration algorithm for registration. Experiment results show that the proposed method is computationally efficient and accurate.
引用
收藏
页码:300 / 305
页数:6
相关论文
共 50 条