臺灣博碩士論文加值系統

隨著人類社會科技水平的提高，對於光學產品的需求與品質也不斷升高，其中光學鏡頭也隨著畫質的要求演變出新的製程技術與光學材料。本文研究對象主要針對小尺寸的光學鏡頭，這些小尺寸的光學鏡頭主要應用於手機的相機模組，為了滿足畫質和成本的需求，在材料上大多選擇塑膠鏡片並透過鏡片的堆疊達到更好的光學性質。由於鏡片缺陷種類多和反光的特性，導致傳統的視覺檢測難以辨識，所以目前對光學塑膠鏡片的檢測主要依賴人工檢測，然而這種方式已難以滿足現代產量大、效率高、一致性強的檢測需求。因此，本研究設計了一個基於機器視覺系統的光學塑膠鏡片缺陷檢測系統。通過對實際生產線問題的實際探討，本文提出了一個基於YOLOv8訓練的檢測模型，並結合移動機構進行影像擷取與辨識的機器視覺系統。該系統旨在提高光學塑膠鏡片表面缺陷的檢測準確性。

With the advancement of human society's technological level, the demand for optical products and their quality continues to rise. Among them, optical lenses have evolved with new manufacturing processes and optical materials to meet the demands for image quality. This study focuses primarily on small.sized optical lenses, which are mainly used in smartphone camera modules. To meet the requirements for image quality and cost, plastic lenses are commonly chosen as materials, and better optical properties are achieved through the stacking of lenses. Due to the variety of lens defects and their reflective characteristics, traditional visual inspection methods struggle to identify them accurately. Therefore, the current detection of optical plastic lens defects relies mainly on manual inspection. However, this approach is increasingly unable to meet the modern requirements for high production volume, efficiency, and consistency in inspection. Hence, this research aims to design and implement a machine vision.based system for detecting defects in optical plastic lenses. Through practical discussions on issues in actual production lines, this paper proposes a detection model based on YOLOv8 training, combined with a mobile mechanism for image capture and recognition in the machine vision system. The system aims to improve the efficiency and accuracy of detecting surface defects in optical plastic lenses, thereby effectively controlling product quality and enhancing detection accuracy.

摘要 ..................i
Abstract ............ii
誌謝 ...............iii
目錄 ...............iv
表目錄 ..................v
圖目錄 ...............vi
符號說明 .....................vii
第一章 緒論 ............1
1.1 研究動機 ...............1
1.2 文獻回顧 .........1
1.3 研究方法 .........3
1.4 研究大綱 ...............4
第二章 鏡片缺陷檢測方法與硬體架構 ...............5
2.1 塑膠鏡片缺陷檢測分析 ...............5
2.2 影像擷取系統 ..................9
2.3 硬體設備規劃 ..................12
第三章 深度學習缺陷檢測方法 ..................14
3.1 YOLOv8模型 ..................14
3.2 深度學習瑕疵分類架構 ..................17
3.3 塑膠鏡片瑕疵圖像數據庫 ..................18
3.4 塑膠鏡片瑕疵檢測訓練與分析 ..................22
第四章 研究結果與分析 ..................24
4.1 線上設備測試 ..................24
4.2 檢測流程的效率 ..................27
第五章 結論 ..................28
參考文獻 ..................29
Extended Abstract ..................30

[1]許佳蓉，2019，”手機鏡頭塑膠鏡片表面瑕疵分類”，國立臺北科技大學自動化科技研究所碩士學位論文。
[2] Redmon, J., Divvala, S., Girshick, R., & Farhadi, A., 2016, "You Only Look Once: Unified, Real- Time Object Detection", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 779–788.
[3] Haiyang T., Shan L., Dan Y., & Yongjie Q., 2023, "A visual defect detection for optics lens based on the YOLOv5 -C3CA-SPPF network model", Optics Express, 31, 2, pp. 2628-2643.
[4] Wang, C., Sun, Q., Dong, X. et al., 2024, "Automotive adhesive defect detection based on improved YOLOv8", SIViP 18, 3, pp. 2583–2595.
[5] Wei, Z. H., Zhang, Y. J., Wang, X. J., Zhou, J. T., Dou, F. Q., & Xia, Y. H., 2024, "A YOLOV8.based approach for steel plate surface defect detection", Metalurgija, 63, 1, pp. 28-30.
[6] Lu, J., Lee, S. H., Kim, I. W., Kim, W. J., & Lee, M. S., 2023, "Small Foreign Object Detection in Automated Sugar Dispensing Processes Based on Lightweight Deep Learning Networks", Electronics, 12, 22, pp, 4621.
[7] Liu, Z., & Ye, K., 2023, "YOLO-IMF: An Improved YOLOv8 Algorithm for Surface Defect Detection in Industrial Manufacturing Field", In International Conference on Metaverse, pp. 15-28
[8] Zhang, L., et al., 2024, "Research on Improved YOLOv8 Algorithm for Insulator Defect Detection", Journal of Real-Time Image Processing, 21, 1, pp. 22.
[9] Li, W., Solihin, M.I., & Nugroho, H.A., 2024, "RCA: YOLOv8-Based Surface Defects Detection on the Inner Wall of Cylindrical High-Precision Parts," Arab Journal of Science and Engineering, pp. 1-19.
[10] Jocher, G., Chaurasia, A., & Qiu, J., 2023, "YOLO by Ultralytics. ", GitHub, Available online: https://github.com/ultralytics/ultralytics.