Uncertainty in interpretation of range imagery
暂无分享,去创建一个
A description is given of a method for automatically deriving a Bayesian probability network from CAD based 3D models. This network may be used to match 3D objects to features derived from range images in the presence of uncertainty. The network describes how evidence for subparts of an object should be combined to calculate the evidence for an object. A prototype model based range interpretation system using this method is also presented. Sample results for an object are presented. For this prototype system, sources of evidence include maximum likelihood estimate fits of circular cylinders, planes, and spheres, as well as boundary data (i.e. edges).<<ETX>>
[1] R. Nevatia,et al. 3-D Surface Description Using Curvature Properties , 1987, Photonics West - Lasers and Applications in Science and Engineering.
[2] Thomas O. Binford,et al. Bayesian inference in model-based machine vision , 1987, Int. J. Approx. Reason..
[3] D. M. Chelberg. An approach to geometric modeling using generalized cylinders and interpretationof range images using bayesian networks , 1990 .
[4] Tod S. Levitt,et al. Utility-based control for computer vision , 2013, UAI.