自動化組裝常需將不同零組件進行裝配,但零組件大小與形狀不同,若零組件未被事先整列而雜亂堆疊,則無法進行自動夾取與組裝,而必須以人工方式拿取與組裝。本研究將使用Intel RealSence D435i景深攝影機來拍攝堆疊之螺絲墊片,判斷位於最上方之螺絲墊片,而由機器手臂夾取。攝影機分別有景深鏡頭以及RGB鏡頭,景深鏡頭檢測螺絲墊片堆疊高度來判斷螺絲墊片堆疊高低。RGB鏡頭將拍攝螺絲墊片進行影像辨識其位置。六軸機器手臂則主要負責移動至影像辨識所得到之位置,再以電動夾爪夾取螺絲墊片進行後續處理,來解決拿取堆疊零件之問題,可用於生產自動化之應用。
Automatic assembly is to assembly different components. However, when components with different sizes and shapes are not lined up and stacked in order, automatic clamping and assembly fails to work. In this study, Intel RealSence D435i depth-of-field camera will be applied in this study to photograph a washer stack for a robot arm to aim at the target washer on the top. The depth-of-field camera has a depth-of-field lens and a RGB lens. With the depth-of-field lens, the height of the screw spacer stack can be measured, while RGB lens helps to identify and locate its exact position for the arm. Based on the information collected from the image recognition, the six-axis robot arm moves to the position of its target washer to pick the target screw washer for subsequent processing. This approach can be adopted to solve the problem of picking up stacked parts, and be further application in production automation.