- 學位論文
應用機器視覺於六軸機械手臂之物件堆疊
Application of Machine Vision with a Six-Axis Robotic Arm for Object Stacking
摘要
近年來產業自動化的比例逐年上升，生產線導入機器視覺取代人類視覺已成為趨勢，利用視覺辨識判斷瑕疵或追蹤夾取物件位置，結合六軸機械手臂取代人力，以解決勞工因為長時間工作下疲勞所造成生產良率降低的問題。本研究採用眼在手(Eye in Hand)架構，於機械手臂末端裝設深度攝影機(RGB-D Camera)用來擷取工作區影像。攝影機初始化會進行相機校正(Camera Calibration)，透過對攝影機擷取之RGB影像做相機校正取得影像畸變誤差值，用以驗證影像畸變程度。手臂與視覺之間的座標轉換關係則透過手眼校正完成。整合系統將RGB影像進行灰階化、自適應二值化、輪廓偵測，並得知物件在影像中X、Y座標值。深度攝影機取得深度資訊，並將深度值轉換機械手臂座標之Z軸向的數值，判斷工作區物件後，結合六軸機械手臂與真空吸盤模組，執行吸取物件以及堆疊物件等工作，以實現自動化堆疊之目的。
並列摘要
In recent years, industrial automation has seen a steady increase. Machine vision has emerged as a popular choice to replace human vision on production lines. It enables defect identification and object tracking, while six-axis robotic arms are integrated to replace human labor. This helps address the issue of reduced production yield caused by worker fatigue due to long working hours. This study adopts the Eye in Hand architecture with a depth camera (RGB-D Camera) installed at the end of the robotic arm to capture working area images. Camera calibration during initialization validates image distortion. Eye-hand calibration establishes the coordinate transformation between the robotic arm and visual system. The integrated system processes RGB images by converting them to grayscale and adaptive threshold and contour detection to determine object X and Y coordinates. The depth camera acquires depth information and converts the depth values into Z-axis coordinates in the robotic arm's coordinate system. After object detection in the working area, the system integrates the six-axis robotic arm with a vacuum suction module to execute tasks such as object grasping and object stacking, enabling the automation of stacking processes.