Grant: Collaborative Research and Development Grants

Dozens to hundreds of data projectors, all working in synchrony, could create quite remarkable information displays, such as at tradeshows, large-scale theatre, theme parks, or other artistic performances. Such a display setup may involve cutting-edge, highly creative projection onto any number of complex, non-flat contexts, such as mapping onto buildings, domes, and other three dimensional objects.The calibration of such projectors is exceptionally difficult, requiring that the pixel coordinate be quite precisely known for each data projector able to illuminate a given location. Currently one or more very high-resolution cameras are mounted on each data projector, to see precisely the illuminated scene for which this data projector is responsible. The cameras are expensive and the process of calibration is slow and time-consuming, significantly limiting the potential markets for such advanced multi-projector installations.For our industrial partner, a significant fraction of their customers would, in fact, be interested in operating projectors in such complex configurations, and more would do so if the tools to enable such configurations were available.In this project we aim to research and develop a novel framework for projector calibration, with goals of reducing cost and time. In particular, could cheaper (lower resolution) cameras be used, but still achieving the necessary accuracy in calibration? As an ambitious extension, the most desirable goal would be to perform system calibration with a single hand-held mobile camera.