Game Development Reference
In-Depth Information
When trying to determine the location and orientation of an actor, there
are three basic techniques: optical, magnetic and mechanical.
Optical
Optical suits use two or more cameras and highly reflective markers on an
actor's suit. The location of the cameras relative to each other is set to a
high degree of accuracy. Motion tracking software follows each marker
from a single camera viewpoint. Motion tracking is a form of pattern
recognition. Software scans the bitmap of a single frame from a video
camera. Then the next bitmap in the video sequence is scanned trying to
locate the same pattern. The location of the marker on the bitmap defines
a ray emanating from the camera, which goes through the marker and
then disappears off to infinity. Using the second and often a third camera,
the marker is tracked from other camera angles. Each camera provides a
line from the camera to infinity. Software can then determine the
intersection of these lines and using the knowledge of the camera world
position provide an x , y , z location in world space for a marker for each
frame of video. At certain times in the video sequences, markers will be
hidden due to the angle of the camera in relation to the performer; this is
described as occlusion. At such times knowledge of the topography of
markers on the performer helps the software to provide a best estimate of
marker location. This type of error checking can make optical motion
capture a very accurate form of analysis. Other benefits of optical capture
are the low cost of additional markers and the lack of restrictions to the
Figure 12.3 Optical motion capture.
Search Nedrilad ::




Custom Search