Center For Visualization & Virtual Environments
Research Project:
Video Rate 3D Data Acquisition using Structured Light
Investigators:
Laurence Hassebrook
Daniel Lau
Chun Guan
Sponsors:
NASA, University of Kentucky
We have developed a scientific method to combine multiple patterns into a single projection image. This allows the subject to move because the camera does not have to synchronize with the projector and can capture at video rate. Since the idea is new, we are systematically improving it.
Our strategy is to first improve the accuracy of the depth reconstruction. We are presently working on projecting the composite pattern using near infrared (NIR) light. This way, the subject will not see the light pattern and a color image will be captured simultaneously with the NIR depth pattern. The next research step is to achieve high lateral and depth resolution from composite patterns. Because there are multiple channels, there are aliasing and band limits that constrain the reconstructed surface accuracy. However, because the bands contain correlated or redundant information we believe it is theoretically possible to restore the original surface topology with high accuracy. A more distant step in the process is to use multiple sensors to capture topologies from different directions that can be combined into one 3D surround topology.
For more information, see: http://128.163.147.208/lgh/research.html