In the paper, the authors used a hybrid tracking system that utilized ARToolKit as a foundation. The hybrid system comes from allowing the PDA to act as a standalone computation device, which can work autonomously, as well as allowing a PC connected wirelessly to shoulder the expensive tracking computation and thus increase overall performance. The PDA was given SoftGL for drawing, which is a light version of OpenGL. Due to the limitations of the PDA being unable to utilize floating points, which OpenGL relies on quite extensively, there were slight performance losses due to translations between integers and floats. Overall, the PDA + camera addon were able to achieve performance of approximately 5 fps when utilizing a supporting PC for computations, and otherwise 2.5-3.5 fps. This is promising for our project, as it demonstrates that even hardware that is not meant to handle augmented reality and has comparatively limited computing limitations can achieve modest results.
Source:
Daniel Wagner; Dieter Schmalstieg; “First Steps Towards Handheld Augmented Reality,” Vienna University of Technology, Favoritenstr.
http://www.icg.tu-graz.ac.at/Members/daniel/Publications/HandheldAR_ISWC03final.pdf
No comments:
Post a Comment