A clinical prototype system for projector-based augmented reality: calibration and projection methods

摘要

In the last two years, the Institute for Process Control and Robotics has developed a prototype system for projector-based augmented reality in cranio-maxillo-facial surgery. The system essentially consists of two CCD cameras and a common video projector. These are used to track the patient, registrate his position (coded light) and to directly project the preoperatively defined planning data onto his surface. System Calibration: While Tsai's method [1] seems to be gold standard, we have developed a new and easy to implement calibration model for cameras and video projectors. The model parameter set consists of the optical center, two vectors that align with the pixel edges, two scalars to describe the location of pixel (0,0) and two scalars to consider radial lense distorsion. The model parameters can easily be found by iteratively solving an overdetermined linear equation system. The necessary point correspondences for the projector are generated by using the coded light phase shift approach both horizontically and vertically. Projection methods: Before the planning data are projected onto the patient's bended surface, an appropriate surface representation has to be calculated. This step is performed each time, the surface significantly changed. Thereby, we defined six projections modes dealing with different visualization requirements (e.g. an access path, viewpoint etc.).