檢測設備的優劣取決於自動對焦系統和檢測系統，而檢測系統之辨識率高低與影像擷取清晰與否有著密切關係，倘若鏡頭焦距調整不佳，將導致無法擷取到最清晰之影像。基於鏡頭焦距對於擷取清晰影像有其重要性，故本研究旨在發展一套具有即時性之自動對焦搜尋演算法。由於自動對焦系統之動態數學模式推導不易，故本研究應用智慧型控制理論之模糊理論設計自動對焦系統之控制器，以影像之清晰度值做為控制器之輸入變數，透過模糊推論的方式預測最佳焦距之位置，期能達到自動對焦之目的。而以往模糊理論之歸屬函數都以試誤法來進行設計，其常常需要耗費大量的時間，且得不到最佳效果。本研究則分別以田口方法和基因演算法來進行歸屬函數之最佳化設計。田口方法係利用直交表進行實驗規劃，進而求出模糊系統歸屬函數之水準回應值，以建立回應表及回應圖，而決定各控制因子之顯著性，將顯著因子用來預測最佳條件之信號雜音比，並以變異數分析法檢驗實驗之誤差，以了解各模糊系統歸屬函數對品質特性之影響，再以確認實驗以檢驗實驗之再現性，進而獲得最佳歸屬函數之組合。而基因演算法具有適應性擇優複製、交配與突變之機制，使得系統在處理問題時不易陷入局部最佳解，而逐步向整體最佳解收斂。並經實務驗證結果顯示，所設計出之自動對焦控制器均可達到快速擷取清晰影像之目的。

The quality of inspection equipment is determined by the auto-focus and inspection systems. The identification rate of inspection system is closely related to the sharpness of image capturing, thus, if the adjustment of lens focal length is poor, it would be impossible to capture the sharpest image. As the lens focal length is critical to capturing sharp images, this study aimed to develop a set of real-time auto-focus search algorithms. As the dynamic mathematical model of an auto-focus system is difficult to derive, this study applied the fuzzy theory of the intelligent control theory in order to design an auto-focus system controller. By using image sharpness as the input variable of the controller, this study applied fuzzy inference to predict the position of the optimal focus points in order to achieve auto-focus. As the membership functions of the fuzzy theory are commonly designed by the trial-and-error method, it is time-consuming and may not obtain the optimal effect. This study employed the Taguchi method and genetic algorithm to achieve the optimal design of membership functions. The Taguchi method uses an orthogonal array for experimental design to obtain the response values of the membership functions of the fuzzy system, and establishes a response table and response graph. The significances of the control factors were determined, and these significant factors were used to predict the signal-to-noise ratio of the optimal conditions. Variance analysis was conducted to verify the errors of the experiment in order to understand the effects of the membership functions on quality characteristics. The repeatability of the experiment was confirmed in order to obtain a combination of the optimal membership functions. Since a genetic algorithm has mechanisms of adaptive preferential replication, crossover, and mutation, the system is unlike to be trapped in a local optimal solution when solving problems, thus, converging toward a global optimal solution. Based on the verification results, the auto-focus controller, as designed by this study, could successfully capture sharp images.

1.Kehtarnavaz, N., and Oh, H. J., Development and Real-time Implementation of a Rule-based Auto-focus Algorithm, Real-Time Imaging, 9, 197-203 (2003).
2.Chen, C. M., Hong, C. M., and Chuang, H. C., Efficient Auto-focus Algorithm Utilizing Discrete Difference Equation Prediction Model for Digital Still Cameras, IEEE Transactions on Consumer Electronics, 52, 1135-1143 (2006).
3.Yu, S., and Mantian, L., Research on SUSAN Based Auto-focusing Algorithm for Optical Microscope Application, 2006 IEEE International Conference on Mechatronics and Automation , 1237-1241 (2006).
4.Nayar, S. K., Watanabe, M., and Noguchi, M., Real-time Focus Range Sensor, Fifth International Conference on Computer Vision, 995-1001 (1995).
5.Widjaja, J., and Jutamulia, S., Wavelet Transform-based Autofocus Camera Systems, IEEE Asia-Pacific Conference on Circuits and Systems, 49-51 (1998).
6.Mahmood, M. T., Choi, W. J., and Choi, T. S., PCA-based Method for 3D Shape Recovery of Microscopic Objects from Image Focus Using Discrete Cosine Transform, Microscopy Research and Technique, 71(12), 897-907 (2008).
7.Zheng, J. Y., Sakai, T., and Abe, N., Guiding Robot Motion Using Zooming and Focusing, IEEE/RSJ International Conference on Intelligent Robots and Systems, 3, 1076-1082 (1996).
8.He, J., Zhou, R., and Hong, Z., Modified Fast Climbing Search Auto-focus Algorithm with Adaptive Step Size Searching Technique for Digital Camera, IEEE Transactions on Consumer Electronics, 49, 257-262 (2003).
9.Jiwen, C, and Jiubin, T., Fast Auto-focusing Based on Partial Image Characteristics, Proceedings of SPIE - The International Society for Optical Engineering, 7155, 1-10 (2008).
10.Lee, J. H., Kim, Y. S., Kim, S. R., Lee, I. H., and Pahk, H. J., Real-time Application of Critical Dimension Measurement of TFT-LCD Pattern Using a Newly Proposed 2D Image-processing Algorithm, Optics and Lasers in Engineering, 46, 558-569 (2008).
11.Yang, M., and Luo, L., A Rapid Auto-focus Method in Automatic Microscope, International Conference on Signal Processing Proceedings, 4697180, 502-505 (2008).
12.Beveridge, G. S., and Schechter, R. S., Optimization: Theory and Practice, McGraw-Hill, New York, (1970).
13.Kesseler, W., and Fischer, J., Analytical Model of Autofocus Systems with CCD Camera, Proceedings of SPIE - The International Society for Optical Engineering, 3965, 396-380 (2000).
14.Chern, N. K., Neow, P. A., Marcelo, H., and Ang, J., Practical Issues in Pixel-Based Autofocusing for Machine Vision, Proceedings of the IEEE International Conference on Robotics and Automation, 3, 2791-2796 (2001).
15.Rahman, M. T., and Kehtarnavaz, N., Real-time Face-priority Auto Focus for Digital and Cell-Phone Cameras, IEEE Transactions on Consumer Electronics, 54, 1506-1513 (2008).
16.Gamadia, M., Peddigari, V., Kehtarnavaz, N., Lee, S. Y., and Cook, G., Real-time Implementation of Auto Focus on the TI DSC Processor, Proceedings of SPIE The International Society for Optical Engineering, 5297, 10-18 (2004).
17.Kehtarnavaz, N., and Oh, H. J., Development and Real-time Implementation of a Rule-based Auto-focus Algorithm, Real-Time Imaging, 9, 197-203 (2003).
18.Jung, J. Y., and Im, Y.T., Fuzzy Algorithm for Calculating Roll Speed Variation Based on Roll Separating Force in Hot Rolling, International Journal of Mechanical Sciences, 42, 249-272 (2000).
19.Wenzhong, Q., Jincai, S., and Yang, Q., Active Control of Vibration Using a Fuzzy Control Method, Journal of Sound and Vibration, 275, 917-930 (2004).
20.Paramasivam, S., and Arumugam, R., Hybrid Fuzzy Controller for Speed Control of Switched Reluctance Motor Drives, Energy Conversion and Management, 46, 1365-1378 (2005).
21.Cheng, C. C., Young, M. S., Chuang, C. L., and Chang, C. C., Fabrication Optimisation of Carbon Fiber Electrode with Taguchi Method Biosensors and Bioelectronics, 18(7), 847-855 (2003).
22.Ali, N., Neto, V. F., Mei, S., Cabral, G., Kousar, Y., Titus, E., Ogwu, A. A., Misra, D. S., and Gracio, J., Optimisation of the New Time-modulated CVD Process Using the Taguchi Method, Thin Solid Films, 469-470, 154-160 (2004).
23.Michael, N., Artificial Intelligence: A Guide to Intelligent System, Addison-Wesley, (2005).
24.Keigo, I., Engineering Optics, Springer Science+Business Media, New York, (2008).
25.Mamdani, E. H., Application of Fuzzy Algorithms for Control of Simple Dynamic Plant, Proceedings of the Institution of Electrical Engineers, 121(12), 1585-1588 (1974).
26.Hsu, S. C., Liu, C. H., and Wang, N. J., Fuzzy PI Controller Tuning for a Linear Permanent Magnet Synchronous Motor Drive, The 27th Annual Conference of the IEEE Industrial Electronics Society, 1661-1666 (2001).
27.Uddin, M. N., Radwan, T. S., Rahman, M. A., and George, G. H., Fuzzy Logic Based Position Control of Permanent Magnet Synchronous Motor, Electrical and Computer Engineering, 1, 93-97 (2000).
28.Holland, J. H., Adaption in Natural and Artificial System, Ann Arbor, University of Michigan Press, (1975).