Finding a sequence of sensor operations for volumetric inspection

A system for exhaustively inspecting a workpiece, not merely certain features, is presented. The system is based on comparing a volumetric representation of a reference object to a volumetric representation of an actual object iteratively created from sensor data. The use of volumetric representations gives this approach a number of distinct advantages over the more traditional boundary-based inspection methods. The method is capable of handling arbitrarily complex geometries and does not rely on the identification of features. It lends itself naturally to a multisensor environment, and facilitates true 3-D inspection as opposed to methods that require dimensional reductions of data by projection onto a 2-D plane. The authors focus on an important aspect of volumetric inspection which is the automatic generation of a sequence of sensor operations that will acquire the complete geometry of the object being inspected.<<ETX>>