Sentience
From AGIRI.org
The Sentience vision project is an attempt to allow machines to perceive the three dimensional structure of their physical environment with reasonable fidelity using inexpensive camera hardware. Started in 2000, it was initially intended to be used as a means of easily creating 3D models of architecture, but was soon diverted for use as a perception system for humanoid robots.
Sentience contains stereo correspondence and grid mapping components which permit distances to objects to be measured and used to generate probabilistic models of spatial occupancy. Using dense correspondence some object recognition tests have been carried out, and autonomous navigation and detailed map building are also planned.
[edit]
