Abstract

Projection mapping [also known as Spatial Augmented Reality] is a well known augmented reality technique which changes the appearance of the physical, real-world environment by projecting media on them. Projection Mapping is currently used in a variety of applications and it is increasingly becoming more popular. In this research, we investigate the use of computer graphics and computer vision techniques that can enable us to create an immersive augmented reality environment in an outdoor setting with large number of participants. Specifically, we address long range projection of stereoscopic media in anaglyph format. This is inexpensive and less intrusive among user immersion technologies, since participants have only to wear low cost anaglyph glasses. Since the projection surface is at a long distance, and is a real world object(s), and there is always ambient lighting, there are many challenges. These include calibration of the cameras and projectors, modeling the projection surface(s) geometry and light behavior and transforming the projection media to compensate for the surface properties. We address these problems and propose an innovative new framework which uses image-based techniques for automatic calibration, geometry and appearance properties capture, and a simulation based technique for creating compensated media. We have implemented our framework and demonstrate its effectiveness through multiple user studies in real world settings.