Please use this identifier to cite or link to this item:
標題: 影像識別技術應用於交換機分機卡瑕疵檢測
PBX Extension Card Surface Defect Inspection by Using Image Recognition Technology
作者: 洪宗毅
Hung, Tsung-Yi
關鍵字: image processing;影像處理;template matching;printed circuit board;defect detection;image recognition;樣板比對;印刷電路板;缺陷檢測;影像識別
出版社: 電機工程學系所
引用: [1] 林仲芬(1993),《影像辨認技術》,全華科技圖書公司,p65~p66。 [2] 鍾國亮(2006),《影像處理與電腦視覺導論》,東華書局,p101。 [3] 張宏林(2008),《Visual c++數位影像模式-識別技術及工程實踐》,文魁資訊股份有限公司,p211。 [4] 林仲芬(1993),《影像辨認技術》,全華科技圖書公司,p4。 [5] 黃春融.詹寶珠(2003),《由影像處理到電腦視覺》,科學發展361期, p12~p19。 [6] 邱奕契.陳易泰(2000),《印刷電路板焊墊尺寸量測之研究》,2000年PCB製造技術研討會,pp110~117。 [7] 陳子超(2003),《表面黏著元件之光學自動檢測》,國立清華大學動力機械工程研究所碩士論文。 [8] 劉權霈(2002),《應用電腦影像視覺技術於PCB自動檢測系統之設計及開發》,國立交通大學工業工程與管理學系研究所碩士論文。 [9] 蘇家興(2005),《印刷電路板表面黏著元件視覺檢測系統》,國立清華大學動力機械工程研究所碩士論文。 [10] 許辰合(2002),《印刷電路板基本元件圖像之萃取系統》,國立成功大學電機工程學系碩士論文。 [11] 陳璋琪(2004),《利用小波理論於印刷電路板缺點之檢測》,國立成功大學電機工程學系碩士論文。 [12] 張維娜(2003),《BGA基板視覺檢測》,國立交通大學電機與控制工程系所碩士論文。 [13] 鄭睿夫(2001),《影像處理技術於載帶規範檢測之應用》,國立成功大學工程科學研究所碩士論文。 [14] R. C. Gonzalez and R. E. Woods(2008),Digital Image Processing,3rd Edition. [15] Milan Sonka(1999), Image Processing ,Analysis ,and Machine Vision.
本篇論文提出了以影像識別技術為基礎,並結合CCD camera應用於交換機分機卡缺陷檢測系統。
電子元件利用表面黏著技術(SMT,surface mount technology)固定於電路板上,有一部份將產生缺陷;另一種情況則是客戶使用過程中,遭受到雷擊或突波衝擊,導致電路板損壞。維修人員進行缺陷修補,通常以人工方式進行檢測,往往浪費許多時間。本研究中探討之方法先將影像灰階化、直方圖均化、二值化、去除多餘雜訊,皆屬於影像處理的部分,再利用輪廓偵測技術使其邊緣之特徵得以擷取,最後利用樣板比對演算法之基礎,進行比對辨識。


In this thesis, we present an image recognition technology combining with the CCD camera for the switch extension card of the defect detection system.
When using SMT (surface mount technology) to fix the electronic components on the circuit board, there will be some defects. Besides, when customers use the switch extension card, the circuit board may suffer from damage due to the impact of lightning or surges. To repair the defects, the detection works are often taken manually, which may lead to the waste of time. In this research, we start from the image processing, including the image grayscale, histogram equalization, thresholding, and noise removal. Then, the techniques of contour edge detection are applied for feature extraction. Finally, we use the algorithms of template matching for matching and recognition.
In our experiment, the images of normal extension card, such as the images of electronic components and solder side, are extracted and used as a standard samples. Meanwhile, the images of the burned IC are put in the database module as a record of abnormal data. To verify the applicability of this detection system, we have tested a set of extension cards and shown significant performance of our research.
其他識別: U0005-2507201021364700
Appears in Collections:電機工程學系所

Show full item record

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.