硬碟片在電鍍後必須將其表面加以拋光再進行後續的加工。然而，在拋光的過程之中硬碟片的表面有時難免會受到外物或是磨粒刮傷，於碟片表面造成刮痕(scratch)。所以拋光之後，需加以檢測，藉由硬碟片表面刮傷的檢測，進而追蹤前段製程有無瑕疵，並加以調整修正，避免造成更多的損失。
本研究目的即在利用數位影像進行硬碟片刮痕檢測。能否從影像中有效檢測出刮痕，將會受光刮痕特性(如方向、大小、深淺)，取像方向，光源方向等因素所影響，因此本研究將探討這些參數的影響，用以設定最佳檢測參數。
論文最後得到下列實驗結論：漫射法比映射法有較佳的辨識效果，刮痕平均角度越大時，有越佳的辨識效果，光照角度80°時有較佳的辨識效果，相機取像角度越接近90°時有較佳的辨識效果，樣品取像角度為270°時有較佳的辨識效果，高解析度時有較佳的辨識效果。採用以上參數組合在刮痕影像辨識將可得到較佳的辨識效果。

The aim of this study is to use digital image analysis to inspect Hard-Disk surface scratches. Hard-Disk surface is usually polished after electroplated. Abrasive particles and other materials are the main cause of the scratches on hard-disk surface during the polishing process. After polishing, hard-disk surface scratches needs to be inspected. To avoid any further loss, defects in the former process should be tracked, regulated and after all, corrected.

Scratch characteristics (ex: direction, size, and depth etc.) will greatly affect the inspection done by digital image. The direction to get the image, the angle of the light, and several other variables would also have effects on the inspection. This study is to discuss the effect of these different parameters and set up the best inspection parameters according to the result.

The result of this study is as the following. Inspection in diffusion is better than in mapping. The larger the estimated average angle of the scratch is, the better the inspection would be. The inspection get to the best when the angle of light source is set at 80 degrees, the angle to get the image at 90 degrees, and the angle to get the image of the object at 270 degrees. Inspections being done in high resolution get better result than it is done in low resolution. Using upward parameters in inspection for hard-disk surface scratches leads to better results, too.

[1] S. C. Lin , and M. L. Wu, “A Study of Effects of Polishing Particle Distributions on Surface Characteristics,” 2000 Proceedings VLSI Multilevel Interconnection Conference (VMIC), Santa Clara, pp. 522-527, June, 2000.

[2] S. S. Chen, “Machine Vision”, ChWa, Taipei, 1997.

[3] S. C. Lin , and H. R. Lu, “A Feasibility Study of Identify Scratch Using Digital Image Analysis,” Proceedings of 2002 Japan-USA Symposium on Flexible Automation, vol.3, Hiroshima, Japan pp. 1019-1022, July, 2002.

[4] Y. N. Sun, “Image Process in Industry Inspection”, Image Process Conference, Kun Shan University of Technology, Taiwan (in Chinese) 1992.

[5] H. Rau, J. W. Tzeng , and S. L. Hsu，“A Study of Defect Inspection of IC Printed Mark”，Journal of the Chinese Institute of Industrial Engineers, Vol. 20, No. 4, pp. 327-336, 2003.

[6] C. H. WU, C. S. HUANG, and H.J. CHEN, “The Recognition and Classification of Defects of Lace”, Master Thesis, National Taipei University of Technology, Taiwan (in Chinese), 2001.

[7] D. M. Tsai, and Y. H. Tseng, “Automatic Surface Inspection for TFT-LCD Array Panels Using Fourier Reconstruction”, AOI Forum & Show 2004, 2004.

[8] A. Branca, W. Delaney, F. P. Lovergine, and A .Distante, “Surface Defect Detection by Texture Analysis with a neural Network”, IEEE International Conference on Robotics and Automation, pp. 1497-1502, 1995.
[9] Ben D. Beake, María Jesús Ibañez García, and James F. Smith, “Micro-impact testing: a new technique for investigating fracture toughness”, Thin Solid Films Volume: 398-399, pp. 438-443, 2001.

[10] T. Nagai, A. Imanishi, and Y. Nakato, “Scratch induced nano-wires acting as a macro-pattern for formation of well-ordered step structures on H-terminated Si(111) by chemical etching”, Applied Surface Science Volume: 237, Issue: 1-4, pp. 533-537, 2004.

[11] M. Kusunoki, Y. Takano, K. Nakamura, M. Inadomaru, D. Kosaka, A. Nozaki, S. Abe, M. Yakoo, and et. al. “Demonstration of surface resistance mapping of large-area HTS films using the dielectric resonator method”, Physica C Volume: 383, Issue: 4, pp. 374-378, 2003.

[12] P. Panjan, M. Cekada, and B. Navinšek, “A new experimental method for studying the cracking behaviour of PVD multilayer coatings”, Surface and Coatings Technology Volume: 174-175, pp. 55-62, 2003.

[13] G. Sorensen, H. Jensen , and U. M. Jensen, “Reactively sputtered Cr nitride coatings studied using the acoustic emission scratch test technique”, Surface and Coatings Technology Volume: 74-75, Issue: 1-3, pp. 297-305, 1995.

[14] P. J. Blau, and W. A. Simpson Jr., “Applications of scanning acoustic microscopy in analyzing wear and single-point abrasion damage”, Wear Volume: 181-183, Part 1;, pp. 405-412, February, 1995.

[15] A. A. Voevodin, R. Bantle, and A. Matthews, “Dynamic impact wear of TiCxNy and Ti-DLC composite coatings”, Wear Volume: 185, Issue: 1-2, pp. 151-157, June, 1995.

[16] R. Myler. Harley, “Fundamentals of machine vision”, Bellingham, Wash.: SPIE Optical Engineering Press, 1999.

[17] Jacob Beck, Barbara Hope, and Azriel Rosenfeld, “Human and machine vision”, New York: Academic Press, 1983.

[18] Milan Sonka, Vaclav Hlavac, and Roger Boyle, “Image processing, analysis, and machine vision”, 1999.

[19] Wesley E., Snyder, and Hairong Qi., “Machine vision”, New York: Cambridge University Press, 2004.

[20] David A. Patterson, and John L. Hennessy, “Computer organization and design: the hardware/software interface”, Amsterdam; Elsevier/Morgan Kaufmann, 2004.

[21] Svelto Orazio, “Principles of lasers”, New York: Plenum Press, 1989.

[22] H.P. Freund , and T.M. Antonsen, Jr., “Principles of free-electron lasers”, London ;Chapman & Hall, New York, 1996.
[23] K. Kanatani, “Geometric computation for machine vision”, Oxford: Clarendon Press; New York: Oxford University Press, 1993