2020 Volume 25 Issue 2 Pages 138-147
Projector calibration is a cumbersome process when building a spatial augmented reality system. Many methods were proposed to simplify the process by using a projector-camera system (PROCAMS), and structured light (SL) projection. Conventional PROCAMS calibration methods use one or more stationary cameras. The camera position needs to be carefully chosen to deal with the problem of trade-off between occlusion and depth error. The camera installation and setup also cost extra labor and time. In this paper, we propose a geometric calibration method using a mobile camera. Unlike conventional methods that use temporal coded SL, we use a spatial coded SL so that it can be decoded in each camera frame. We take advantage of the multiview images of the projection surface, and use two types of image features, to achieve a robust calibration result. Experiments show that the result of our method is comparable with that of a checkerboard-based approach.