Numerical Solution of the Four-Dimensional Nonstationary Fokker-Planck Equation资料.pdf

Handling Uncertainty in 3D Object Recognition Using Bayesian Networks B. Krebs, M. Burkhardt and B. Korn emaih{B.Krebs, M.Burkhardt, B.Korn}~tu-bs.de Institute for Robotics and Computer Control, Technical University Braunschweig, Hamburger Str. 267, D-38114 Braunschweig, F.R.G. A b s t r a c t . In this paper we show how the uncertainty within a 3d recog- nition process can be modeled using Bayesian nets. Reliable object fea- tures in terms of object rims are introduced to allow a robust recognition of industrial free-form objects. Dependencies between observed features and the objects are modeled within the Bayesian net. An algorithm to build the Bayesian net from a set of CAD models is introduced. In the recognition, entering evidence into the Bayesian net reduces the set of possible object hypotheses. Furthermore, the expected change of the joint probability distribution allows an integration of decision reasoning in the Bayesian propagation. The selection of the optimal, next action is incor- porated into the Bayesian nets to reduce the uncertainty. 1 I n t r o d u c t i o n The task to perform recognition for a 3d scene is a highly complex process which involves various types of sensor processing and interpretation algorithms. A hu- man observer identifies the objects by the knowledge of what he expects to see. If an unknown object occurs the observer may try to find a clue by looking from different directions or may try to detect specific features to solve the ambiguity. Thus, the cognitive process of recognition should