A Bayes nets-based prediction/verification scheme for active visual reconstruction

International audienceWe propose in this paper an active vision approach for performing the 3-D reconstruction of polyhedral scenes. To perform the reconstruction we use a structure from controlled motion method which allows a robust estimation of primitive parameters. As this method is based on particular camera motions, perceptual strategies able to appropriately perform a succession of such individual primitive reconstructions are proposed in order to recover the complete spatial structure of complex scenes. Two algorithms are proposed to ensure the exploration of the scene. The former is a simple incremental reconstruction algorithm. The latter is based on the use of a prediction/verification scheme managed using decision theory and Bayes Nets. It allows the visual system to get a more complete high level description of the scene. Experiments carried out on a robotic cell have demonstrated the validity of our approach