｜曲面物件瑕疵檢測簡介

摘要：本研究探討一種曲面物件瑕疵檢測視覺系統，該視覺檢測系統具有下列兩項特性：（1）可克服不同視角所產生的表面反光問題；（2）可避開反光之視角進行三維表面瑕疵檢測。本技術使用影像濾波處理分離影像中正常光源之影像與直接反射光源所造成的反光影像，實驗結果顯示本系統可有效避免不同視角所造成的反光區域，相較於傳統的二維視覺檢測系統只可用於平面檢測（例如晶圓或電路板），本曲面物件檢測系統可實際應用於工業產品的三維檢測上且不需精確的校正，並可避開導致反光的視角以最少的取像視角涵蓋整個檢測面達到檢測的行為。

Abstract: The purpose of this paper is to develop a vision system for surface defect detection of 3D objects. It aims at the ill-defined qualitative items such as stains and scratches. A robot vision system for surface defect detection may counter challenges such as （1） high surface reflection at some viewing angles and （2） no reference markers in any sensed images for matching. A filtering process is used to separate the illumination and reflection components of an image. Tests were performed on a variety of hand-held electronic devices such as cellular phones. Experimental results show that the proposed system can reliably avoid reflection surfaces in different viewing angles. Traditional visual inspection systems mainly work for 2D planar surfaces such as PCBs and wafers. The results have practical applications for industrial objects with arbitrary surfaces under image misalignment for 3D objects. The proposed system can identify the viewing angles with minimum surface reflection.

關鍵詞：機器視覺、瑕疵檢測、三維曲面

Keywords：Machine vision, Defect detection, 3D surface

緣起
由於手攜式電子產品多為曲面或高反射性表面，其他如鋼珠表面亦具相同特性，此等目標物表面容易因有角度光源投射至表面反射而形成光影或陰影，如圖1所示。圖1 （a）為手機外殼光滑金屬表面反光情形，圖1 （b）則為車頭燈表面光影現象。當以視覺系統做自動化檢測時，瑕疵特徵反而受限於曲面角度、反光及對比度不足導致時常無法突顯，因此視覺辨識技術困難度極高，非一般平面靜態影像檢測系統可克服。加上面對未來少量且多變的3C產品線，產品皆為曲面高反光材質，在自動外觀瑕疵檢測上目前並無一個較佳的解決方案，因此本研究將著手開發此類型的三維物件曲面瑕疵檢測系統。