BGA (Ball Grid Array) 可提供較好的電氣與熱性能，因此成為晶片製程上的一種普遍封裝技術。為了辨別每顆晶片的型號與功能，雷射刻印在BGA封裝表面上刻上圖案。良好的刻印，必須仰頼精確的校準。本論文提出一個次像素BGA晶片校準方法，透過影像檢測系統來量測刻印封裝IC表面的中心位置與偏向角。

本論文提出的方法，首先在檢測的區域內，利用區域投影法產生橫向與縱向的投影曲線來辨識物體的輪廓，藉此輪廓定義影像興趣區域。然後使用DOB濾波器檢測出物體的邊緣點。此外，對於物體邊緣有裂痕或破裂的像素，則用邊緣追隨法濾除。接下來，為了達到次像素的精確度，本論文使用一內差函數次像素邊緣檢測方法來修正邊緣點。最後，經由最小平方差直線匹配的迴歸演算以及幾何運算，得到校準中心位置與偏向角。

本論文利用人工合成影像來評估所提方法的穩定性、精確度、及執行效能。在沒有雜訊的干擾下實驗結果顯示中心定位誤差少於正負0.07像素，偏向角誤差少於正負0.01像素;在不同程度的高斯雜訊干擾及人工合成影像下，誤差標準差少於1%。在一台Pentium4 2.8 GHz處理器的電腦上，完成一晶片的校準需要時間約41毫秒，平均檢測範圍為 像素。

BGA (Ball Grid Array), a technology of packaging, is widely used in the manufacturing process of integrated circuits because it can offer better electrical and thermal performance. In order to identify the model name and function of each IC chip, the laser marking is used to generate patterns on the surface of the BGA package. But good marking relies on accurate alignment for IC package. In this thesis, a subpixel alignment method for BGA images is proposed to measure the central position and orientation angle of an IC package for the location of the etched mark through an image inspection system.

In the proposed method, region projection is first applied to generate vertical and horizontal projection profiles and recognize the contour of the object within the inspected area and delimit the regions of interest (ROI) by the projection profiles. Then the DOB (Difference of Boxes) filter is used to detect edge elements of the object. Further, the edge elements resulting from flaws or cracks on object’s boundaries are filtered out by the edge following method. Next, in order to achieve sub-pixel accuracy, a deterministic sub-pixel edge detection method using interpolation functions is used to modify the edge elements. Finally, the central position and orientation angle are obtained by the LSE line fitting algorithm followed geometric computation.

Our proposed method is evaluated in terms of the stability, accuracy and running time for synthetic images. The experimental result for noiseless synthetic images show that the error of the central position and orientation can be less than 0.07 pixel and 0.01 degree, respectively; for synthetic images under various levels of Gaussian noise corruption, the error standard deviation of the central position and the orientation are all less than 1%. The running time is measured with a Pentium 4 processor of 2.8 GHz, it takes approximately 41 milliseconds to complete the entire the alignment process for a inspection area in our sample images.

[1] F. Bouchara, “Efficient Algorithm for Computation of the Second-order Moment of the Subpixel-edge Position,” Applied Optics, Vol. 43, No.23 , 2004.
[2] J. F. Canny, “A Computation Approach to Edge Detection,” IEEE Trans. on Pattern Anal. Machine Intell., Vol. PAMI-8, No. 6, pp. 697-698, Nov. 1986.
[3] D. Demigny, “On Optimal Linear Filtering for Edge Detection,” IEEE Trans. on Image Processing, Vol. 11, No. 7, pp. 728-737, Jul. 2002.
[4] F. Devernay, “A Non-maxima Suppression Method for Edge Detection with Sub-pixel Accuracy,” INRIA Sophia Antipolis Research Report, No. 2724, 1995.
[5] S. Ghosal and R. Mehrotra, “Detection of Composite Edges,” IEEE Trans. on Image Processing, Vol. 3, No. 1, pp. 14-25, 1994.
[6] R.C. Gonzalez and R. E. Woods, Digital Image Processing, Edition, Prentice-Hall, NJ, 2002.
[7] H. S. Hou and H. C. Andrews, “Cubic Splines for Image Interpolation and Digital Filtering,” IEEE Trans. Acoust., Speech, Signal Processing, Vol. ASSP-26, No. 6 pp. 508–517, Dec. 1978.
[8] A. Hrertas, G. Medioni, “Detection of Intensity Changes with Subpixel Accuracy Using Laplacian-Gaussian Masks.” IEEE Trans. Vol. PAMI-8, No. 5, pp. 651-664, 1986.
[9] A. A. Kassim, H. Zhou and S. Ranganath, “Automatic IC Orientation Checks,” Machine Vision and Applications, Vol. 12, pp. 107-112, 2000.
[10] R.G. Keys, “Cubic Convolution Interpolation for Digital Image Processing,” IEEE Trans. on Acoustics, Speech and Signal Processing, Vol. ASSP-29, No. 6, pp. 1153–1160, Dec. 1981.
[11] W. C. LEE, “Subpixel BGA Images Alignment Based on Moment Methods,” Master Thesis, Dept. of Electrical Engineering, NCKU, June 2005.
[12] Y. Li, T. Young, and J. Margerl, “Subpixel Edge Detection and Estimation with a Microprocessor-Controlled Line Scan Camera.” IEEE Trans. on Industrial Electronics, Vol. 35, No. 1, pp. 105-112, 1988.
[13] D. Marr and E. Hildreth, “Theory of Edge Detection,” Proceeding of the Royal Society of London, Vol. B 207, pp. 187-217, 1980.
[14] V. S. Nalwa and T. O. Binford, “On Detecting Edges,” IEEE Trans. on Pattern Anal. Machine Intell., Vol. PAMI-8, No. 6, pp. 669-714, Nov. 1986.
[15] F. A. Pellegrino, W. Vanzella and V. Torre, “Edge Detection Revisited,” IEEE Trans. on Systems, Man, And Cybernetics-Part B: Cybernetics, Vol. 34, No. 3, pp. 1500-1518, Jun. 2004.
[16] M Petrou and J. Kittler, “Optimal Edge Detectors for Ramp Edges,” IEEE Trans. on Pattern Anal. Machine Intell., Vol. 13, No. 5, pp. 483-491, May. 1991.
[17] Y. D. Qu, C. S. Cui, Chen, and J.Q. Li, “A Fast Subpixel Edge Detection Method Using Sobel-Zernike Moments Operator,” Image and Vision Computing, Vol. 23, Issue1, pp. 11-17, Jan. 2005.
[18] S. Randani, F. Bouchara, and K. Djemal, “Probability Distribution of Subpixel Edge Detection,” Signal Process. Vol. 84, pp. 445-452, 2004.
[19] S. Sarkar and K. Boyer, “On Optimal Infinite Impulse Response Edge Detection Filters,” IEEE Trans. on Pattern Anal. Machine Intell., Vol. PAMI-13(11), Nov. 1991.
[20] Y. Shan, G.W. Boon, “Sub-pixel Location of Edges with Non-Uniform Blurring: a Finite Closed-Form Approach,” Image and Vision Computing, Vol. 18, pp. 1015-1023, 2000.

[21] M. Sonka, V. Hlavac and R. Boyle, Image Processing Analysis, and Machine Vision, Edition, P.W.S, 1999.
[22] C. H. Tao, “A Study on Alignment of BGA Images for Automatic Visual Inspection,” Master Thesis, Dept. of Electrical Engineering, NCKU, June 2004.
[23] X Wang, J Gao, L Wang, “A Survey of Subpixel Object Localization for Image Measurement,” Proceedings of International Conference on Information Acquisition, pp. 398-401, 2004.
[24] J Ye, G Fu, and UP Poudel, “High-Accuracy Edge Detection with Blurred Edge Model,” Image and Vision Computing, Vol. 23, pp. 453-467, 2005.