DATA FUSION FROM MULTIPLE CAMERAS FOR AUTOMATIC DISASSEMBLY

摘要

In this paper we analyze, in some detail, the vision system architecture for disassembly applications. This work is carried out in the context of motion and stereo analysis. The methodology presented is useful for work in manufacturing conditions facing difficult situations like the occlusion of components. The recognition and location of three-dimensional objects is important for automatic disassembly. A data fusion from a multiple-camera scheme has been proposed for extracting information from the scene. Data provided by some sensors is used to determine object recognition, location and orientation.