抄録
In this work we combine optical flow and stereo vision algorithms, for obtaining a three dimensional object tracking system capable two follow changes varying in time. The optical flow algorithms allows the study of displacements varying in time and by using a stereo vision system we can retrieve the coordinates of the objects observed in the three axes (x, y, z) and therefore the reconstruction of such objects. A stereo vision system was implemented in Matlab using two identical webcams placed in parallel geometry. The ultimate goal of this project is to have a portable and affordable system that takes into account the disadvantages intrinsic of the implementation such as limited resolution and distortion presented on it for being webcams. Experimental results and system limitation analysis will be presented.
