本論文提出將SIFT 導入自動光學檢測系統的方法，以提升自動光學檢測系統對物件相對位置及角度的容忍度，期望能提高自動光學檢測的準確性及降低系統的架設成本。在本論文研究中主要以SIFT的尺度不變性及旋轉不變性來定位待測物的位置及角度，以增加影像比對的準確性，並取代自動光學檢測系統中以機械手臂或輸送帶將待測物品定位至指定位置及角度的步驟，進而降低自動光學檢測系統的架設成本。 This study proposed a method of importing Scale-invariant feature transform (SIFT) to Automatic Optical Inspection System (AOIS) to improve the tolerance of object position and rotation detection in AOIS, to increase the precision and decrease the implementation cost. This study uses the scale & rotational invariant of SIFT to locate the position and the angle of object to increase the precision of image matching, and replace the step of locating the object to designated position with robotic arm and conveyor of AOIS. And decrease the AOIS implementation cost.
