Stereo computation combining structured light and passive stereo matching

In this work we propose a stereo computation method which initially borrows structured light strategy based on single phase shifting approach. An accurate phase shifting code allowed us to considerably decrease the candidate set of points compared to passive stereo matching. Then, once the most similar match was found using area based matching, we re-use the accurate wrapped code to refine the initially found disparity value. Our method is extremely simple to implement and in that sense very promising for real time applications. Shown comparison results demonstrate that we can produce accuracy comparable with the state-of-the art methods in stereo matching.