在工廠的生產線上,要進行較複雜的加工任務或物件取放任務時,機械手臂通常會導入電腦視覺系統,藉以輔助自動化產線,以提高其效率,而本論文旨在研究與開發一種基於視覺之自動化物件取放流程。 本論文使用二指平行夾爪發展出一套自動夾取流程,針對具角點的任意角柱形物體,結合影像處理並設計演算法,提供一個簡易的使用者介面,讓使用者能夠快速且有效率地執行夾取工作。此一流程首先使用一固定位置之相機拍攝物體,接著對輸入影像進行影像處理後可以獲得角柱截面的所有角點位置;再來,利用這些角點位置計算最佳之夾持點,最後命令機械手臂移動至夾持點位置並夾取物體。本研究並實際測試了多種角柱形物體,以證實本夾取流程的可行性。
In the production line of a factory, to perform complex machining work or object pick-and-place, a computer vision system is commonly added to assist the automated production line so as to improve the efficiency. The purpose of this thesis is to develop a vision-based automated object pick-and-place planning. This thesis developed an automated object pick-and-place plan using two-fingered gripper. The plan, suitable for polygonal prisms with identifiable corner points, integrated image processing and design algorithm into a robotic system and provided a user interface to allow users to manipulate the robot to grip the object in a quick and efficient way. In the last, this work validated the process with a real robot.