利用機器視覺偵測鑄件毛邊與瑕疵

Use Machine Vision to Detect the Burrs and Defects of Casting Part

盧文鈞(Wen-Jean Lu)

指導教授 ： 蔡篤銘

元智大學/工程學院/工業工程與管理學系/碩士(1992年)

摘要

本論文主要在於利用機器視覺技術來偵測鑄件所產生的毛邊和瑕疵。本論
文所探討的鑄件將以平面(二度空間)工件為主，並且鑄件外形須為平滑曲
線(Smooth Curve)。而機器視覺所要偵測的部分為產生於接模面的不規則
毛邊和由於鑄造過程不當所造成的瑕疵。本研究可適用於沒有任何特徵
點(如內孔、邊角等)的鑄件。本研究的具體成果包括：1)能偵測出鑄件毛
邊同時並自動檢測鑄件之邊緣形狀是否有瑕疵，以確保品質，2)不利用特
徵點(如內孔、邊角等)為主要方法，所以即使毫無特徵點或是特徵點完全
為毛邊所覆蓋，也能正確地找出毛邊的所在處。

關鍵字

機器視覺 ； 鑄件毛邊 ； 鑄件瑕疵

並列摘要

Burrs(excess material) are usually squeezed out around the
periphery of the mold. Poor control of foundry process may
result in a casting with a peripheral breakdown. Those excess
of unwanted metal and peripheral defects are irregular and vary
considerably from casting to casting. This research aims at the
development of a machine vision process for automatic detection
of burrs and peripheral crumble of a casting. The non-contact
detection result can be used for automatic visual inspection of
casting's peripheral shape. And the detected form and location
of burrs can be further applied to automatic deburring system.
This research concentrates on two-dimensional flat casting,
which can be represented by a smooth, closed curve. Special
feature points such as interior holes of a casting are not
required to carry out the recognition process.

並列關鍵字

Machine Vision ； Burrs Detection ； Peripheral Defect Inspection