摘要

This paper describes grasp point selection on an item of clothing randomly placed on a table. The input data for our proposed method are a set of range images captured from a fixed, 3D range camera. Hem elements are extracted from the data, and their relationships are characterized for both similarity measures and grasp point evaluation. Experiments using real images, targeting a piece of clothing, show the effectiveness of the proposed method.

  • 出版日期2015-1-2