No abstract

No abstract

Today's world automobile market is highly competitive. Quality of automobile is considered even more important and thus efforts are being made to improve the quality of automobile parts in order to survive in such competitive market. Despite rapid development of inspection devices using the vision system, there are few inspection systems for automobile parts. The reason is because the inspection with only one camera is nearly impossible, since automobile parts are three-dimensional and large. The seat frame, one of the typical automobile parts, has various models and is assembled using more than 20 subparts. This paper suggests a flexible assembly inspection system of seat frame through the following procedures: first, the inspection method is standardized so that the system can cope with not only some currently existing models but also latent models in the future. Second, 6 cameras are used for inspecting seat frame assembly; 4 cameras are for measuring accuracy, and 2 for right and left-hand sides. Third, a simple pattern matching method is proposed for checking the position of assembly parts. Finally, slit beam method is proposed for checking the existence of assembly parts. Especially, it is very effective to inspect assembly of automobile parts with 3D shape and complex background image.

No abstract

The generation of a volume swept by a tool along its path, called the swept volume, is indispensable for the volumetric simulation of five-axis machining. In some cases of five-axis machining, the swept volume may be self-intersecting, which requires special processing to handle the geometric and topological problems due to complex tool motion. In this work, an algorithm is proposed for generating the swept volume, including the self-intersection case for five-axis ball-end milling, based on triangle strip representation. The proposed algorithm has various applications such as interference detection in assembly design, and visualization of manipulator motions in robotics.