A Vision-Measuring Method for an Imperfect Circle

影像量測瘕疵圓之方法

10.29977/JCIIE.200501.0005

李廣齊(Kuang-Chyi Lee)；林見昌(Chien-Chang Lin)

眾數法 ； 圓心定位 ； 瘕疵圓 ； 影像量測 ； mode method ； center positioning ； imperfect circle ； vision measurement

工業工程學刊

22卷1期（2005 / 01 / 01）

38 - 45

英文

在半導體元件的製程中，印刷電路板(PCB)的圓孔定位以及光學點所提供焊接時座標的參考依據，常因檢測時應打光的方式有所不同或是半導體元件的金屬表面反光，而造成CCD攝影機在取像時圓孔變形、產生雜訊或是光線不足導致影像模糊，其使得在求取定位圓心的座標時產生誤差。本文提出一四點量測圓心的眾數法，以克服上述瘕疵圓影像量測的問題。由模擬與實際測試的結果知此方法可行，其並且可以準確量測到被遮蔽六十度以上圓之圓心座標。本文亦將四點量測圓心之眾數法和面積法、三點法、輪廓追蹤法與最小方差法等影像量測圓心的方法比對，四點量測圓心之眾數法實為較快速且又準確的圓心量測方法。

The positioning center of the circular hole of PCB and the reference coordinates of welding by optical point in the manufacture process of the semiconductor components will cause the circle deformed and will create image noise or blurred image when the image from CCD camera is grabbed by the different lighting method or light source. When the surface of objects inspected reflect the light, it will cause the error of center positioning. This paper proposes a 4-points mode number method for the center positioning to overcome the problem of the measurement of the imperfect circle. The results of simulation and experiments indicate that this method is feasible and it can measure accurately the coordinates of the center of the circle which is covered by the other objects or is deformed over sixty degrees. Based on experiments, it is known that the 4-points mode method is faster and more accurate than the area, 3-points, contour-tracking and least-squares error methods of center-positioning.

1. Ahn, S. J.,W. Rauh,H.J. Warnecke(2001).Least-squares orthogonal distances fitting of circle, sphere, ellipse, hyperbola and parabola.Pattern Recognition,34,2283-2303.
2. Atherton, T. J.,D. J. Kerbyson(1999).Size invariant circle detection.Image and Vision Computing,17,795-803.
3. Aviad, Z.,E. Lozinskii(1987).Semantic thresholding.Pattern Recognition Letters,5,321-328.
4. Brink, A. D.(1989).Grey level thresholding of images using a correlation criterion.Pattern Recognition Letters,9,335-341.
5. Forbes, A. B.(1988).Least-squares best-fit geometric element.NPL Rep DITC,37,503-506.
6. Guil, N.,E. L. Zapata(1997).Lower order circle and ellipse Hough transform.Pattern Recognition,30(10),1729-1744.
7. Haddon, J. F.(1988).Generalized threshold selection for edge detection.Pattern Recognition,21(3),195-203.
8. Huang, J.(1999).An exact solution for the roundness evaluation problems.Precision Engineering,23,2-8.
9. Jywe, W. Y.,C. H. Liu,C. K. Chen(1989).The min-max problem for evaluating the form error of a circle.Measurement,26,273-282.
10. Lee, K. C.,H. P Huang,S. S. Lu(1992).A flexible robot inspection cell based on vision image.International Journal of Computer Integrated Manufacturing,5(5),310-318.
11. Lee, K. C.,H. P. Huang,S. S. Lu(1993).Burr detection by using vision image.International Journal of Advanced Manufacturing Technology,8,275-284.
12. Luo, D.,P. Smart,J. E. S. Macleod(1995).Circular Hough transform for roughness measurement of objects.Pattern Recognition,28(11),1745-1749.
13. Murthy, T. S. R.(1986).A comparison of different algorithms for circularity evaluation.Precision Engineering,8,19-23.
14. Murthy, T. S. R.,S. Z. Abdin(1980).Minimum zone evaluation of surface.International Journal of Machine Tool Design and Research,20,123-136.
15. Nakagawa, Y.,A. Rosenfeld(1978).A note on the use of local min and max operations in digital picture processing.IEEE Transaction on System, Man and Cybernetics,8(8),632-635.
16. Pei, S. C.,J. H. Horng(1995).Circular arc detection based on Hough transform.Pattern Recognition Letters,16,615-625.
17. Ridler, T. W.,S. Calvard(1978).Picture thresholding using an iterative selection method.IEEE Transaction on System, Man and Cybernetics,8(8),630-632.
18. Ventura, J. A.,S. Yeralan(1989).The minimax center estimation problem for automated roundness inspection.European Journal of Operational Research,41,64-72.
19. Weszka, J. S.,R. N. Nagel,A. Rosenfeld(1974).IEEE Transaction on Computers.
20. Zuniga, O. A.,R. M. Haralick(1988).Gradient threshold selection using the facet model.Pattern Recognition,21(5),493-503.