Teleoperation and augmented reality

Abstract

We have developed a depth map transform algorithm that converts the depth data according to a video image. The algorithm consists of four steps: a depth map analysis, a video image analysis, a depth map and video image relation creation and finally a depth surface transformation depending on the relation established previously. We show how these four steps are performed assuming that the two maps have nearly equal features. This technic is used with nearly compatible depth map and video image in order to create a quite compatible set allowing virtual objects integration in real images. This is an important aspect of augmented reality since it enables the hidden faces calculation of virtual objects in coherence with the video image even when work conditions hinder accurate camera calibration or model of the real scene is not perfectly known.