3-D image processing and grasping planning expert system for distorted objects

Abstract

Generally, the conventional pattern matching method is hard to apply to objects which have individually variable sizes, such as ceramic products, because of their expansion, shrinkage, or cracking in the baking process. To solve this problem, especially for automatic manipulation of these kinds of products, the authors propose an expert system. Avoiding individual differences, the system reasons about and recognizes bending, twists, and cracks of an object by integrating image information about six surfaces. The system calculates and reasons about the most suitable grasping point, given the manipulator's working area and the position of the object. When the most suitable grouping point is not in the working area, the grasping planning mode is activated. The effectiveness of this method was confirmed experimentally.<>