臺灣博碩士論文加值系統

中文摘要
一物體進行檢測時，需先行利用取像設備取得物體影像，再對影像進行二值化處理。通常一張影像是由所謂的像素(Pixel)所構成，像素的灰階值介於0~255之間，代表著視覺上的強度變化，為影像基本計算單位。在一般工業上應用自動化視覺檢測，影像處理是一個必須之階段，因為須檢測之物體大部份都是由直線及圓弧組成的簡單型物體，當影像擷取後，再針對所獲取之影像進行二值化轉換，以便對物體進行辨識。
本研究是利用Ker所提出之K-曲率法則，計算並找出二維物體邊界上分斷點，再將其拋棄即可精確找出邊界上之線性方程式，再以Borland C++ Builder 程式語言撰寫比對程式即可進行實際微鑽頭影像分析及檢測出待測物之端面分離與否結果，主要係利用CCD取得之微鑽頭刀刃影像，再將其可分為影像預處理、鑽頭邊界搜尋、運用斷點尋找線性方程式及影像檢測等四個部份，主要內容包括：(1) 影像預處理(2) 邊界搜尋(3)曲率計算(4)斷點計算(5)辨別分離(合格或不合格)。

Abstract
When an object is under examination, firstly it has to obtain the image of the object by using the acquisition facility binary image, and then the two-dimension process can be taken. Usually the image is composed of pixel. The value of the gray scale is between 0~255, which is the basic measurement unit that indicates the visual effect of different tints. In general automatic visual examination that applies in the industry, image process is a necessary procedure. The examined items are mostly simple objects that composed of lines and curves. After the image is being taken, the two dimension process can be conducted to identify the object.
The subject research applies the K-curvature Rules to measure and identify the break point over the edge of the two-dimension object. When the break point is identified, desert the break point to get the accurate linear equation .Continue to compose the contrast equation n the language of Borland C++ Builder, and then it is ready to proceed with actual micro-drills image analyze and to examine the result of edge gap. It utilizes the micro-drills knife-edge image acquired by CCD. The process can be divided into four parts: The major tasks includes (1)image processing (2)boundary description (3) curvature compute (4) break point compute (5) recognition gap

Keyword：two-dimension process、pixel、K-curvature Rules、curvature、the break point

目 錄

致謝辭……………………………………………………………………I
中文摘要……………………………………………………………………II
Abstract……………………………………………………………………III
目錄……………………………………………………………………IV
圖目錄……………………………………………………………………V
表目錄……………………………………………………………………VI
第一章 緒論……………………………………………………………………1
1.1研究動機與背景…………………………………………………………1
1.2研究目的…………………………………………………………………5
1.3研究方法與步驟…………………………………………………………7
1.4研究範圍及限制………………………………………………………9
第二章 文獻探討……………………………………………………………10
2.1視覺影像檢測………………………………………………………… 10
2.2 Ker之斷點檢測法則……………………………………………………13
2.2.1 K曲率法之基本定…………………………………………………13
2.2.2 臨界值計算……………………………………………………… 15
2.2.3數位化後產生誤差之修正……………………………………… 15
2.3微鑽頭不良之探討…………………………………………………… 18
第三章 研究方法……………………………………………………21
3.1研究方法與流程……………………………………………………21
3.2系統流程說明……………………………………………………22
3.3演算步驟說明………………………………………………30
第四章 實例驗證……………………………………………………………34
4.1實驗過程與結果………………………………………………………34
4.2綜合討論……………………………………………………………44
第五章 結論與建議…………………………………………………………46
5.1結論……………………………………………………………………46
5.2未來研究方向…………………………………………………………46
參考文獻……………………………………………………………………… 48

參考文獻D. M. Tsai, Boundary-based corner detection using neural networks, Pattern Recognition, 30(1), 85-97(1997)D. M. Tsai, H. T. Hou, H. J. Su, Boundary-based corner detection using eigenvalues of covariance matrices, Pattern Recognition Letters 20, 31-40(1999)F.Leymarie, and M.D.Levine, Curvature Morphology, Center for Intelligent Machines McGill University,Montreal,Canada Dec (1988)Gregory Dudek and John K. Tsotsos, Shape Representation and Recognition from Multiscale Curvature, Computer Vision and Image Understanding Vol.68.November,pp170-172, (1997)H.Freeman and L. S. Davis, A corner-finding algorithm for chain-coded curves, IEEE Trans. Comput. C-26, 297-303 (1977)H. T. Sheu and W. C. Hu, A rotationally invariant two-phase scheme for corner detection, Pattern Recognition 29(5), 819-828 (1996)J. Y. Chiu, Automated Inspection of IC Flaws Based on Computer Vision Methods, Dissertation, National Yunlin University of Science and Technology (2000)J. S. Park, J. H. Han, Contour matching: a curvature-based approach, Image and Vision Computing 16, 181-189 (1998)Kwanghoon Sohn, Winser E.Alexander, Curvature Estimation and Unique Corner Point Detection for Boundary Representation. IEEE Interational Conference on Robotics and Automation Nice,France.May 1992K.E.Lepeshenkov, and E. V. Tchepine, Description and Recognition of Objests Forms Using Nonlinear Terms Basis, Proceedings of the Workshop on Computer Science and Information Technologies CSIT (2000)Rong-Jian Chen and Jui-Lin Lai, Novel approach and system for automatic PCB visual inspection using hybrid flaw detection, Journal of National Llen-Ho Institure of Technology Vol.17.pp1 (2000)Sam Anand, Christopher McCordin and Rohit Sharma An Integrated Machine Vision Based System for Solving the Non-Convex Cutting Stock Problem Using Cenetic Algorithms Journal of Manufacturing Systems Vol.18/No.6(1999)Studies on the Inspection of Mango Fruits by Machine Vision, National Chia-Yi Institute of Agriculture 43, 39-50(1995)Tsorng-Lin Chia, and Win-Jun Cherng, A Parallel Algorithm for Computing Curvatures of Object Contours in a Binary Image, Journal of c.c.i.t Vol27,NO2,Jan, (1998)Vladimir Kovalevsky, Curvature in Digital 2D Images, Pattern Recogonition (2000)Yi-Hong Lin, and Hong-Jane Kaun , Aided by Computer Vision ,Bulletin of National Pingtung Polytechnic Polytechnic Istitute 6(1) pp19-27 (1997)陳時新, 視覺三次元量測系統, 量測資訊第41期 pp52-54 (1995)吳文明,及張俊隆, 測頭字動檢測系統, 機械工業雜誌 二月號pp156-160, (1995)俞有華,及張鴻義 螺帽邊裂自動揀測系統之探討,南開學報第二期 ,pp117-127 (1992)萬一怒 及 楊智超, 電腦類神經網路應用於糙米品質檢測之研究, 農林學報46(4), 59-81(1997)官志忠, 智慧型鑽頭檢測系統,台灣科技大學機械所, (2000)柳美鈴 ,使用K曲率法則於二維物體之分斷點偵測之研究 , 朝陽科技大學工工所, (2000)吳成柯,戴善榮,程湘君,雲立實, 數位影像處理, 儒林圖書公司(1996)