車用鏡面瑕疵自動化檢測系統

An Automated Optical Inspection System for the Defect Detection of Car Mirrors

王燕語(Yen-Yu Wang)

指導教授 ： 章 明

中原大學/工學院/機械工程研究所/碩士(2013年)

https://doi.org/10.6840/cycu201300691

摘要

本研究係針對車用鏡片製程中玻璃表面與反射面瑕疵之自動化同步檢測需求，以高像素線型CMOS擷取影像方式，開發一套生產線上直接執行瑕疵檢測的自動化光學檢測系統，以解決目前生產線仍以人工檢視為主流所面臨的產品良率不穩定及廠商人事成本日增的困境。本研究使用的CMOS解析度為7μm，搭配0.7倍鏡頭後之光學解析度為10 μm，使用側向打光方式將光源投射在車用鏡面樣本上，利用玻璃折射與鏡面反射原理執行瑕疵檢測，在無瑕疵的部分，因光線直接反射，CMOS不會接收到光訊號而呈現黑屏；在鏡面有瑕疵的部分，光線會因漫射而進入CMOS，因而可透過擷取亮點方式檢測出瑕疵位置及判斷瑕疵種類。
本研究利用Microsoft Visual C# 語言撰寫人機介面，執行瑕疵檢測上之應用，其可依照檢測解析度需求與試體尺寸設計檢測系統，較大的樣品可以數個CMOS達到同步檢測的目標，並可以適用於所有類似鏡面玻璃製品的生產線上的檢測應用，實驗結果顯示目前系統的誤判率為0.0325%，瑕疵檢測率為92.4808%。

關鍵字

自動化光學檢測 ； 鏡面瑕疵檢測 ； 線型CMOS

並列摘要

In order to improve the yield and reduce the man-power cost for car mirror industry, an in-line automated optical inspection system has been developed in this study to instead of traditional human inspections. The inspection system consists of a high-pixel number linear CMOS sensor, a lateral lighting source, and a high-resolution translational stage. The pixel size of the sensor is 7 μm. An imaging module with a magnification of 0.7 was adopted in this system, therefore the imaging pixel size was roughly 10 μm. The image of car mirror was observed by scanning the specimen with the help of the translational stage. Since the light source was laterally projected onto the surface of specimen, the dark background was observed for the flawless region due to the refraction and reflection of light, and the defect region can be highlighted due to the scattering of light.
Experimental results show that the defects on both the glass surface and mirror surface of car mirrors produced in the car mirror manufacturing process can be synchronously recognized with a self-developed software. The current defect detection rate is 92.4808% and the false alarm rate is 0.0325%. Larger specimen can be inspected through adding either the number of pixels of CMOSs or the imaging modules with a parallel computing technique. In addition, the system can be applied to the in-line monitoring of production lines for all similar mirror-glass products.

並列關鍵字

Automated Optical Inspection ； Flaw Detection ； Car Mirror ； linear CMOS

參考文獻

[19]黃智政，「應用區塊離散餘弦轉換搭配輝預測模式檢測車用晶面玻璃表面瑕疵」，碩士論文，私立朝陽科技大學工業工程與管理系，台中，2009年。

[7]劉宜興，「光電三維追蹤器之校正自動化研究」，碩士論文，國立成功大學機械工程研究所，台南，2004年。

[14]洪崇祐，「應用一維傅立葉分析於TFT-LCD液晶顯示面板之瑕疵檢測」，碩士論文，元智大學工業工程與管理研究所，桃園，2004年。

[17]Perng , D. B., Chou , C. C., Chen , W. Y., “A Novel Vision System for CRT Panel Auto-inspection” Journal of the Chinese Institute of Industrial Engineers, Vol. 24, No. 5, pp. 341-350, 2007.

[2]Reiss, G., Vancea, J., Wittmann, H., Zweck, J., Hoffmann, H., “Scanning Tunneling Microscopy on Rough Surface: Tip-Shape-Limited Resolution” J. Appl. Phys, Vol. 67, Issue. 3, pp. 1156-1159, 1990.