Journal of South China University of Technology (Natural Science Edition) ›› 2007, Vol. 35 ›› Issue (5): 41-45.
• Electronics, Communication & Automation Technology • Previous Articles Next Articles
Zhao Xin-Can Zuo Hong-fu
Received:
Online:
Published:
Contact:
About author:
Supported by:
国家"863" 计划项目(2006AA04Z427) ;国家自然科学基金资助项目(60672164)
Abstract:
In the augmented reality registration algorithm of fiducial points , many artifical markers are added in the real scene , which makes the system veηcomplex. In order to solve this problem , a robotic visual servo system is analyzed , and an augmented reality registration algorithm based on natural feature with virtual visual servo technology is proposed. Then , by taking the geometrical characters including point and moment as the image characters ,matrices describing the interaction between the variation of image character and that of virtual relative pose , namely , the Jacobian matrices of the image , are presented , by which a 3D registration algorithm based on natural feature is achieved. As compared with the hybrid tracking and the coplane four-point registration algorithms , the proposed algorithm is of higher registration accuracy.
Key words: augmented reality, 3 D registration, natural feature, virtual visual servo
Zhao Xin-Can Zuo Hong-fu. Augmented Reality Registration AIgorithm ßased on Natural Feature[J]. Journal of South China University of Technology (Natural Science Edition), 2007, 35(5): 41-45.
0 / / Recommend
Add to citation manager EndNote|Reference Manager|ProCite|BibTeX|RefWorks
URL: https://zrb.bjb.scut.edu.cn/EN/
https://zrb.bjb.scut.edu.cn/EN/Y2007/V35/I5/41