本研究的檢測項目為引腳剪切成型與經過蓋印的DIP IC元件，利用影像處理技術針對在其上的油墨印刷圖案進行檢測的工作。與IC封裝外觀相同，油墨上印刷圖案也扮演著功能辨識的角色，任何人可以憑藉著IC所有外觀進而判斷它的性質與使用性等。就以印字圖案來說，吾人首先針對影像定位介紹許多方法，接著在印字瑕疵檢測部分同樣提出若干種方法進行分析。
影像定位部份，重點在於定位點的決定，可分為近似直線與近似點，近似直線的部分通常以物件的輪廓點當定位點，如IC膠體的邊緣輪廓點近似，近似點通常是計算出一區域的代表點當定位點，如物件的形心或相關係數搜尋匹配最高的點位置等。
印字瑕疵檢測部分，本研究介紹三種檢測方式，且瑕疵種類因不同檢測方式而有所不同，吾人因此根據印字的方式與業界的建議，重新定義IC印字瑕疵種類，並根據這些瑕疵種類發展一套適合的IC印字瑕疵檢測系統，希望能取代人力成為主流。

The inspect item of this research is DIP IC components which had formed leads and marked characters, make use of image process to inspect ink patterns printed on it. Same as the IC outline packaging, ink printed pattern gives user to identify the function, we can depend on all of the IC's outline to judging properties and usages etc. Focus on marking patterns, this thesis introduced a lot of image alignment methods, and then other marking defect inspection methods.
Image alignment, emphasis on locations of the alignment point, can be divided to line fitting and point fitting two parts. Line fitting's alignment points usually base on contour of the object, like IC mold's contour line fitting. On the other hand, Point fitting's alignment point is calculated on one area, like centroid of the object, or correlation searching for the highest matching point.
Marking defect inspection, this research bring in three inspect methods , and different defect components have different inspect manners, thus according to marking manners and industry's suggestions to define new IC marking defect components, from that to develop proper IC marking defect inspection system, substitute labor and become a tendency.

[1]Intel,“1998 Packaging Databook,”1998.
[2]曾健維,“晶片印字瑕疵檢測之研究(A Study of Defect Inspection of IC Printed Mark),”私立中原大學工業工程學系碩士論文,1999.
[3]A.A. Kassim, H. Zhou, S. Ranganath,“Automatic IC orientation checks,” Department of Electrical & Computer Engineering, National University of Singapore,2000.
[4]邱振洋,“應用電腦視覺於IC 瑕疵檢測之研究(Automatic Inspection of IC Flaws Based on Computer Visions),” 國立雲林科技大學工業工程與管理研究所碩士論文,2000.
[5]劉孝忠, “Microcanonical平均退火類神經網路在晶片印字辨識之應用,”私立中華技術學院機械工程系,2001.
[6]Bernard C. Jiang, Szu-Lang Tasi, and Chien-Chih Wang,“Machine Vision-Based Gray Relational Theory Applied to IC Marking Inspection,” IEEE TRANSACTIONS ON SEMICONDUCTOR MANUFACTURING, VOL. 15, NO. 4, NOVEMBER 2002.
[7]洪欽文,“機器視覺在積體電路板自動化檢驗之應用(Applying Machine vision to the Automated Inspection of IC Boards),”私立義守大學工業工程與管理學系碩士論文,2002.
[8]林祥璋,“應用共變異矩陣法於IC雷射打印瑕疵檢測之研究(Study of Defect Inspection of IC Laser Marking by Using Covariance Matrix Method),”國立台灣科技大學自動化工程及控制研究所碩士論文,2002.
[9]周志賢,“應用次像素插值方法改善印刷電路板缺陷直接檢測之可行性評估,”國立清華大學動力機械工程學系碩士論文,2003.
[10]黃志榮,“自動檢測之定位、量測與辨識之研究,”國立成功大學電機工程學系碩士論文,2003.
[11]戴慶輝,“晶片電阻與標籤字元辨識系統之開發,”國立屏東科技大學機械工程學系碩士論文,2003.
[12]林宜弘,林宜賢和曾皇逢,“IC印碼檢測系統之開發,”屏東科技大學機械系,正修技術學院電子系,AOI Forum & Show 2004.
[13]陳金聖,“LCD驅動IC自動檢測系統,”國立台北科技大學自動化科技研究所研究計畫,2005.
[14]Kamalakannan Ganapathy, Christopher G.Fernando and Asad DavariFast,“Character Recognition System Using Expert Systems,”WVU Tech,2005 IEEE.