Robot Assisted Material Handling for Shirt Collar Manufacturing-Turning andPressing-Volume 5: Three-Dimensional Machine Vision

摘要

Apparel manufacturers are interested in applying advanced automation techniquesto gain increased productivity for exiisting labor-intensive garment assembly processes. The objective of this thesis is to investigate the usefulness of three-dimensional vision sensing for fabric material manipulation. A three-dimensional (3-D) range finder composed of a charge coupled device (CCD) camera and a facility for projecting laser stripes has been developed and implemented in a robotic workstation dedicated for turning and pressing shirt collars. Through image processing and triangulation, the range finder develops position information which is used by the robot to properly position the collar on a pressing surface. The scanning and modeling of a collar takes between 16 and 18 seconds, and the complete sensing and positioning procedure requires 20 seconds with robot velocities of 20 ips. The range finder is capable of locating a 3-D point in space to within a +2 mm precision in the depth direction, and to within ji mm in the orthonormal directions. Collars are positioned on the pressing work surface with a precision of 14 mm which is within tolerance for a successful collar placement. This application of 3-D vision sensing for flexible fabric manipulation demonstrates the promising potential of 3-D vision sensing for apparel manufacturing.