檢測金屬、玻璃或光澤表面塗裝等鏡面材質上的瑕疵，因其光學特性，容易將光源與相機等不相關之物體一併拍攝進影像中，造成後續電腦處理之負擔，也較容易出現誤判，使得系統可靠度下降。此外，部分瑕疵僅於特定入射光角度下才較明顯，因此為此類待測物設計照明方式時，硬體可調整的幅度也會更加侷限。上述解法之一，是Paulo M.F. Forte提出之瑕疵影像增強法，該法以螢幕作為光源，並於螢幕上變換圖案，使得無需活動機構即可達成照明變化，接著藉由演算法將多張取像影像合併，生成一張高瑕疵對比度之最終影像。本研究中提出一補償光源方案，可進一步優化前述瑕疵影像增強法。實驗結果顯示，本方案能有效降低使用瑕疵影像增強法造成之背景暗紋問題，且可減少系統取像張數。

Detecting flaws on mirror materials such as metal, glass or glossy surface coating, due to its optical characteristics, it is easy to capture light sources and irrelevant objects such as cameras into the image, resulting in the burden of subsequent computer processing and prone to misjudgment, making the system reliability decrease. In addition, some defects are only obvious under certain incident light angles, so when designing lighting methods for this type of device under test, the adjustable range of the hardware will be more limited.One of the above solutions is the defect image enhancement method proposed by Paulo M.F. Forte. This method uses the screen as the light source, and changes the pattern on the screen, so that the illumination change can be achieved without moving the mechanism, and then multiple images are captured by the algorithm The images are merged to produce a final image with high defect contrast. In this study, a compensation light source scheme is proposed, which can further optimize the aforementioned defect image enhancement method. Experimental results show that this solution can effectively reduce the problem of background dark lines caused by the use of flawed image enhancement methods, and can reduce the number of images captured by the system.

[1]網頁資源：黃仲宏，產業技術分析，AOI技術發展的動力來自於精密檢測的電子產業需求。
取自： Doit經濟部技術處網站。
https://www.moea.gov.tw/MNS/doit/industrytech/IndustryTech.aspx?menu_id=13545&it_id=411
[2]網頁資源：機器視覺介紹 ∕ AOI介紹
取自：apisce公司網頁：
http://apisc.com/aoiintroduction.htm?fbclid=IwAR3wf03OviyV8DpZ_tOpWxAyPtNbKxSwbgJzD5mQP4C2_M_h6L6uxJQmNa8
[3]張凱翔，「表面黏著型發光二極體光學檢測系統開發」，國立中山大學，碩士論文，2012年
[4]郭丞諴，「效方法應用於零件遺漏偵測及玻璃氣泡檢出」，國立高雄應用科技大學，碩士論文，2013年。
[5]LU, Rong Sheng, et al. AOI techniques for surface defect inspection. In: Applied mechanics and Materials. Trans Tech Publications Ltd, 2010. p. 297-302.
[6]張晴晴，「側照式SMD-LED瑕疵自動光學檢測系統」，國立交通大學，碩士論文，2010年。
[7]陳宗達，「CMOS玻璃蓋片自動光學檢測機台之設計及開發」，國立交通大學，碩士論文，2005年。
[8]林義祥，「機器視覺應用於照相手機鏡片模組邊緣瑕疵檢測」，國立交通大學，碩士論文，2011年。
[9]網頁資源：AOI光源
取自：URVISION公司網頁：
https://www.urvision-tw.com/article_detail/14.htm
[10]黃子航，「玻璃透鏡瑕疵之自動化光學檢測」，國立雲林科技大學，碩士論文，2021年。
[11]陳韋吉，「透明玻璃之自動化可視瑕疵檢測」，朝陽科技大學，碩士論文，2011年。
[12]鄭旭宏，「車用後視鏡之自動化輪廓瑕疵檢驗及尺寸量測」，朝陽科技大學，碩士論文，2013年。
[13]Forte, P.M.F.; Felgueiras, P.E.R.; Ferreira, F.P.; Sousa, M.A.; Nunes-Pereira, E.J.; Bret, B.P.J.; Belsley, M.S. Exploring combined dark and bright field illumination to improve the detection of defects on specular surfaces. Opt. Lasers Eng. 2017, 88, 120–128.
[14]OTSU, Nobuyuki. A threshold selection method from gray-level histograms. IEEE transactions on systems, man, and cybernetics, 1979, 9.1: 62-66.
[15]SAXENA, Lalit Prakash. Niblack’s binarization method and its modifications to real-time applications: a review. Artificial Intelligence Review, 2019, 51.4: 673-705.
[16]蕭家振，「CCD數位相機漏光現象之研究與改善」，國立交通大學，碩士論文，2012年。
[17]HAIN, Rainer; KÄHLER, Christian J.; TROPEA, Cam. Comparison of CCD, CMOS and intensified cameras. Experiments in fluids, 2007, 42: 403-411.
[18]MAGNAN, Pierre. Detection of visible photons in CCD and CMOS: A comparative view. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 2003, 504.1-3: 199-212.
[19]蘇彥熒，「使用於金氧半影像感測器之平行處理遞迴式12位元類比數位轉換器」，國立成功大學，碩士論文，2005年
[20]王龍豪，「高反光物體影像處理技術研究」，國立雲林科技大學，碩士論文，2018年
[21]PAIN, Bedabrata, et al. "A back-illuminated megapixel CMOS image sensor. " 2005.
[22]Sukegawa, Shunichi, et al. "A 1/4-inch 8Mpixel back-illuminated stacked CMOS image sensor." 2013 IEEE International Solid-State Circuits Conference Digest of Technical Papers. IEEE, 2013.
[23]NIE, Kaiming, et al. "A single slope ADC with row-wise noise reduction technique for CMOS image sensor. " IEEE Transactions on Circuits and Systems I: Regular Papers, 2020, 67.9: 2873-2882.
[24]吳星穎，「以機器視覺做陶瓷基板參數量測」，國立雲林科技大學，碩士論文，2021年
[25]網頁資源： Basler ace acA5472-17uc
取自：Basler公司網頁：
https://www.baslerweb.com/en/products/cameras/area-scan-ameras/ace/aca5472-17uc/
[26]VELICHKO, Sergey. Overview of CMOS Global Shutter Pixels. IEEE Transactions on Electron Devices, 2021, 69.6: 2806-2814.
[27]LIANG, Chia-Kai; CHANG, Li-Wen; CHEN, Homer H. Analysis and compensation of rolling shutter effect. IEEE Transactions on Image Processing, 2008, 17.8: 1323-1330.
[28]劉彥醇，「CMOS滾動式快門接收器的可見光通訊傳輸和解調性能改進之研究」，國立交通大學，碩士論文，2018年
[29]ALLEN, Elizabeth; TRIANTAPHILLIDOU, Sophie. The manual of photography. CRC Press, 2012.
[30]LYON, Richard F. Depth of field outside the box. Online Monograph, 2006, 39.
[31]網頁資源：Imaging Fundamentals
取自：Edmund Optics (EO)公司網頁：
https://www.baslerweb.com/en/products/cameras/area-scan-ameras/ace/aca5472-17uc/
[32]網頁資源：TUSS 12MP Series 工業鏡頭
取自：中和碁電公司網頁：https://www.jidien.com/zh_tw/product/ins.php?index_m1_id=33&index_id=802
[33]網頁資源：ASUS VW225T
取自：ASUS公司網頁：https://www.asus.com/tw/SupportOnly/VW225T/HelpDesk_Manual/