Automatic and Continuous Projector Display Surface Estimation Using Everyday Imagery

Projector-based display systems have been used in computer graphics for about as long as the field has existed. While projector-based systems have many advantages, a significant disadvantage is the need to obtain and then adhere to an accurate analytical model of the mechanical setup, including the external parameters of the projectors, and an estimate of the display surface geometry. We introduce a new method for the latter—for continuous display surface autocalibration. Using a camera that observes the display surface, we match image features in whatever imagery is being projected, with the corresponding features that appear on the display surface, to continually refine an estimate for the display surface geometry. In effect we enjoy the high signal-to-noise ratio of "structured" light (without getting to choose the structure) and the unobtrusive nature of passive correlation-based methods. The approach is robust and accurate, and can be realized with commercial off-the-shelf components. The method can be used with a variety of projector-based displays, for scientific visualization, trade shows, entertainment, tele-immersion, or the Office of the Future. And although we do not demonstrate it in this paper, we have also been working on extending the method to include continual estimation of other system parameters that vary over time. CR