臺灣博碩士論文加值系統

光學鏡頭是我們日常生活中眾多設備和系統的基本組件。光學鏡頭的完美功能對其所採用的光學系統的品質和精確度至關重要，確保光學鏡頭符合嚴格的品質標準至關重要。本論文提出了一種透過先進的影像增強技術和深度學習自動進行光學鏡片缺陷檢測的創新方法。研究分為兩個關鍵部分：影像增強和缺陷檢測系統。影像增強著重於優化影像品質。利用 YCbCr 色彩空間的 Y 頻道可大幅減少處理時間，進而增強後續影像處理。使用不銳化掩模可以有效放大對比度，而雙邊濾波器可以成功消除雜訊。這些綜合增強功能可提高缺陷偵測系統中的影像清晰度並降低雜訊。利用霍夫圓變換來偵測光學透鏡的邊界，從而能夠驗證缺陷是否位於光學透鏡內。採用能量分析、大津閾值處理和輪廓檢測來識別和提取缺陷輪廓。深度學習模型用於對缺陷進行分類，幫助製造商輕鬆做出產品品質決策。此缺陷檢測系統即使對於細小、微弱的缺陷也能達到100%的高檢出率，分類率高達97%。此外，每個產品的平均處理時間僅2.5秒。

Optical lenses are fundamental components in a myriad of devices and systems that are part of our daily lives. The flawless functionality of an optical lens is paramount to the quality and accuracy of the optical systems in which it is employed, ensuring that optical lenses meet stringent quality standards is of utmost importance. This thesis presents an innovative approach to automating defect inspection in optical lenses through advanced image enhancement techniques and deep learning. The research is divided into two key components: image enhancement and defect inspection system. Image enhancement focuses on optimizing image quality. Leveraging the Y channel of the YCbCr color space significantly reduces processing time, enhancing subsequent image processing. Employing an unsharp mask effectively amplifies contrast, while a bilateral filter successfully eliminates noise. These combined enhancements lead to improved image clarity and noise reduction, in the defect inspection system. The hough circle transform is utilized to detect the boundaries of the optical lens, enabling verification of whether defects are located within the optical lens. Energy analysis, Otsu's thresholding, and contour detection are employed to identify and extract defect contours. Deep learning models are used to classify defects, helping manufacturers easily make decisions about product quality. This defect inspection system achieves a high detection rate of 100% even for small and faint defects and a classification rate of 97%. In addition, the average processing time per optical lens product is only 2.5 seconds.

CHAPTER 1 ORGANIZATION 1
1.1 Motivation 1
1.2 Dissertation Organization 2
CHAPTER 2 INTRODUCTION 3
2.1 Overview of the Optical Lens Samples 4
2.2 Image Acquisition System 5
2.3 Optical Lens Defect Image 11
2.4 Structure of Optical Lens Defect Inspection System 12
2.5 Summary 13
CHAPTER 3 IMAGE ENHANCEMENT FOR OPTICAL LENS 14
3.1 Image Enhancement Flowchart 14
3.2 YCbCr Color Space 16
3.3 Unsharp Mask 17
3.4 Bilateral Filter 21
3.5 Result of Image Enhancement 24
3.6 Summary 25
CHAPTER 4 DEFECT INSPECTION SYSTEM FOR OPTICAL LENS 26
4.1 Defect Detection 26
4.1.1 Hough Circle Transform 26
4.1.2 Energy Analysis 29
4.1.3 Otsu’s Threshold Method 30
4.1.4 Contour Detection 32
4.2 Defect Classification 34
4.2.1 Types of Defects in Optical Lenses 34
4.2.2 Methods for Defect Classification 35
4.3 Defect Inspection System Flowchart 37
4.4 Summary 39
CHAPTER 5 EXPERIMENTAL RESULT AND DISCUSSION 40
5.1 System Configuration 40
5.2 Defect Detection Experiment 41
5.3 Defect Classification Experiment 44
5.4 Summary 46
CHAPTER 6 CONCLUSION AND FUTURE WORK 47
6.1 Conclusion 47
6.2 Future Work 49

[1]Y-Q. Lin, Y-C. Xiang, Y-X. Lin, and J. Yu, “Defect detection system for optical element surface based on machine vision,” 2019 2nd International Conference on Information Systems and Computer Aided Education (ICISCAE), pp. 415-418, Dalian, China, Sept. 2019.
[2]S. Roy et al., "Deep Learning for Classification and Localization of COVID-19 Markers in Point-of-Care Lung Ultrasound," in IEEE Transactions on Medical Imaging, vol. 39, no. 8, pp. 2676-2687, Aug. 2020.
[3]H. Zhu, X. Wang and R. Chen, "Deep Learning Algorithm Based Remote Sensing Image Classification Research," 2023 IEEE 3rd International Conference on Electronic Technology, Communication and Information (ICETCI), pp. 1368-1373. Changchun, China, 2023.
[4]G. Lou and H. Shi, "Face image recognition based on convolutional neural network," in China Communications, vol. 17, no. 2, pp. 117-124, Feb. 2020.
[5]K. Kim, S. Kang, J. Kim, J. Lee, J. Kim and J. Kim, "Multiple objects recognition for industrial robot applications," 2013 10th International Conference on Ubiquitous Robots and Ambient Intelligence (URAI), Jeju, Korea (South), 2013, pp. 257-259. 2013.
[6]S. -H. Lee, C. -S. Yang, T. -W. Hou and C. -H. Yeh, "An image preprocessing method for fingernail segmentation in microscopy image," 2017 IEEE 2nd International Conference on Signal and Image Processing (ICSIP), pp. 489-493. Singapore, 2017
[7]K. -w. Lee, Y. -s. Kim, S. -h. Park, Suryanto and S. -j. Ko, "Effective color distortion and noise reduction for unsharp masking in LCD," in IEEE Transactions on Consumer Electronics, vol. 54, no. 3, pp. 1473-1477, August, 2008.
[8]Murinto, S. Winiarti, D. P. Ismi and A. Prahara, "Image enhancement using piecewise linear contrast stretch methods based on unsharp masking algorithms for leather image processing," 2017 3rd International Conference on Science in Information Technology (ICSITech), pp. 669-673. Bandung, Indonesia, 2017.
[9]P. D. Patil and A. D. Kumbhar, "Bilateral Filter for Image Denoising," 2015 International Conference on Green Computing and Internet of Things (ICGCIoT), pp. 299-302. Greater Noida, India, 2015.
[10]C. Xiong, L. Chen and Y. Pang, "An Adaptive Bilateral Filtering Algorithm and its Application in Edge Detection," 2010 International Conference on Measuring Technology and Mechatronics Automation, pp. 440-443. Changsha, China, 2010.
[11]F. Porikli, "Constant time O (1) bilateral filtering," 2008 IEEE Conference on Computer Vision and Pattern Recognition, pp. 1-8, Anchorage, AK, USA, 2008.
[12]N. Tuyen Le, J. -W. Wang, M. -H. Shih and C. -C. Wang, "Novel Framework for Optical Film Defect Detection and Classification," in IEEE Access, vol. 8, pp. 60964-60978, 2020.
[13]D. J. Kerbyson and T. J. Atherton, "Circle detection using Hough transform filters," Fifth International Conference on Image Processing and its Applications, 1995., pp. 370-374. Edinburgh, 1995.
[14]Le, N.T, Wang, J.-W, Wang, C.-C, Nguyen, T.N. Automatic Defect Inspection for Coated Eyeglass Based on Symmetrized Energy Analysis of Color Channels. Symmetry 2019, 11, 1518.
[15]N. Otsu, "A Threshold Selection Method from Gray-Level Histograms," in IEEE Transactions on Systems, Man, and Cybernetics, vol. 9, no. 1, pp. 62-66, Jan. 1979.
[16]Xin-Yi Gong, Hu Su, De Xu, Zheng-Tao Zhang, Fei Shen and Hua-Bin Yang. An Overview of Contour Detection Approaches. International Journal of Automation and Computing, vol. 15, no. 6, pp. 656-672, 2018.
[17]Satoshi Suzuki, Keiichi Abe, “Topological structural analysis of digitized binary images by border following”, Computer Vision, Graphics, and Image Processing, vol. 30, pp. 32-46, Issue 1,1985.