三維掃描器的發展日漸成熟，許多領域開始嘗試使用點雲，從自走車的避障與定位、地圖建模，建築物的三維建模等等；到後來嘗試利用到醫療、微觀的檢測，如IC板的瑕疵或是加工刀具的磨耗等等，銑刀的磨耗狀況與加工產品的精度相關，又因其特殊的幾何形狀導致影像視覺檢測上也難以觀測其完整的磨耗狀況，故本論文想藉由整合銑刀不同角度的三維資訊建立銑刀的三維模型，以模型觀察銑刀的整體狀況。本論文首先透過顯微鏡取得銑刀的微觀三維資訊，再利用SIFT(Scale Invariant Feature Transform)演算法找出相鄰平面影像中共同的特徵點並結合高度資訊提出特徵點雲，將其利用SAC-IA(The Sample Consensus-Initial Alignment)粗對齊演算法與ICP(Iterative Closest Point)精對齊演算法計算兩座標系的幾何轉換關係，並轉換銑刀點雲，疊合於全局的模型點雲中，直到加入所有角度資訊後即完成模型建立，最終透過比較將模型投影於平面所得到的外徑尺寸與顯微鏡所得到的實際外徑尺寸，來觀察模型建立的狀況。本研究以Visual C++程式語言、OpenCV、PCL函式庫，實現三維模型的建立。

The development of 3D scanner has progressed over time, and many fields have begun to try to use point cloud data. In large-scale applications, such as obstacle avoidance and positioning functions of AGV, mapping the factory’s 3d information, three-dimensional modeling in the construction industry, etc., it has begun to try to use it in the subsequent fine-dimensional scales, such as the medical field, microscopic defect detection, such as IC board Defects or wear of machining tools, etc. The wear condition of the milling cutter is related to the accuracy of the processed product, and because of its special appearance, it is difficult to observe its complete wear condition in the image visual inspection. Therefore, this paper intends to establish the milling cutter's performance by integrating the three-dimensional information of the milling cutter from different angles. Three-dimensional model to observe the overall condition of the milling cutter. This paper first obtains the microscopic three-dimensional information of the milling cutter through a microscope, and then uses the SIFT (Scale Invariant Feature Transform) algorithm to find the common feature points in the adjacent planar images and combines the height information to propose a feature point cloud, which is used by SAC-IA (The Sample Consensus-Initial Alignment) rough alignment algorithm and ICP (Iterative Closest Point) fine alignment algorithm calculate the geometric conversion relationship between the two coordinate systems, and convert the milling cutter point cloud, superimposed on the global model point cloud, until After adding all the angle information, the model is created. Finally, compare the outer diameter size obtained by projecting the model on the plane with the actual outer diameter size obtained by the microscope, we can observe the quality of the model creation. This research uses Visual C++ programming language, OpenCV, and PCL library to create a three-dimensional model.

[1] M. Pollefeys, "Self-calibration and metric 3D reconstruction from uncalibrated image sequences," PhD thesis, ESAT-PSI, KU Leuven, 1999.
[2] I. Z. Ibragimov and I. M. Afanasyev, "Comparison of ROS-based visual SLAM methods in homogeneous indoor environment," in 2017 14th Workshop on Positioning, Navigation and Communications (WPNC), 2017: IEEE.
[3] Ford’s High-Tech Three-Dimensional Automated Paint Defect Detection System Delivers Maximum Quality. Available: https://www.ford.co.za/about-ford/newsroom/2020/fords-high-tech-three-dimensional-automated-paint-defect-detection-system-delivers-maximum-quality/
[4] P. Tang, D. Huber, B. Akinci, R. Lipman, and A. J. A. i. c. Lytle, "Automatic reconstruction of as-built building information models from laser-scanned point clouds: A review of related techniques," vol. 19, 2010.
[5] 陳正益, "端銑刀磨耗線上量測," 碩士, 機械工程系研究所, 國立中正大學, 嘉義縣, 2018.
[6] 廖伯翰, "端銑刀磨耗監測與壽命預測技術之研發," 碩士, 機械工程系研究所, 國立中正大學, 嘉義縣, 2016.
[7] Z. You, H. Gao, L. Guo, Y. Liu, and J. J. W. Li, "On-line milling cutter wear monitoring in a wide field-of-view camera," vol. 460, 2020.
[8] A. Varghese, V. Kulkarni, S. S. J. J. o. M. S. Joshi, and Engineering, "Tool life stage prediction in micro-milling from force signal analysis using machine learning methods," vol. 143, 2021.
[9] W. Zeng, X. Jiang, and L. J. T. I. J. o. A. M. T. Blunt, "Surface characterisation-based tool wear monitoring in peripheral milling," vol. 40, 2009.
[10] Z. Zhao, X. Liu, C. Yue, R. Li, H. Zhang, and S. J. A. S. Liang, "Tool Quality Life during Ball End Milling of Titanium Alloy Based on Tool Wear and Surface Roughness Models," vol. 10, 2020.
[11] L. Fernández-Robles, L. Sánchez-González, J. Díez-González, M. Castejón-Limas, and H. J. N. Pérez, "Use of image processing to monitor tool wear in micro milling," vol. 452, 2021.
[12] 林哲宇, "使用影像特徵與點雲訊息建構物件模型," 碩士, 機械與機電工程學系碩士班, 淡江大學, 新北市, 2019.
[13] V. Sequeira, J. G. Gonçalves, and M. I. Ribeiro, "High-level surface descriptions from composite range images," in Proceedings of International Symposium on Computer Vision-ISCV, 1995: IEEE.
[14] 吳載根, "線上自動光學檢測系統," 碩士, 機械工程系, 國立臺灣科技大學, 台北市, 2019.
[15] 林俊佑, "基於輪廓線與SIFT 3D 特徵之即時三維人臉點雲建模與辨識系統," 碩士, 資訊工程系研究所, 國立臺北科技大學, 台北市, 2017.
[16] M. Szydłowski, B. Powałka, M. Matuszak, and P. J. P. E. Kochmański, "Machine vision micro-milling tool wear inspection by image reconstruction and light reflectance," vol. 44, 2016.
[17] R. Pieper and A. J. A. O. Korpel, "Image processing for extended depth of field," vol. 22, no. 10, 1983.
[18] L. Čerče, F. Pušavec, J. J. J. o. M. S. Kopač, and Technology, "3D cutting tool-wear monitoring in the process," vol. 29, no. 9, 2015.
[19] J. Tian et al., "Three-dimensional reconstruction of laryngeal cancer with whole organ serial immunohistochemical sections," vol. 10, 2020.
[20] Keyence VHX-7000.
[21] OpenCV. Available: https://opencv.org/
[22] Y. Y. Schechner and N. J. I. J. o. C. V. Kiryati, "Depth from defocus vs. stereo: How different really are they?," vol. 39, no. 2, 2000.
[23] A. R. Weeks, Fundamentals of electronic image processing. SPIE Optical Engineering Press Bellingham, 1996.
[24] S. J. C. v. Suzuki, graphics, and i. processing, "Topological structural analysis of digitized binary images by border following," vol. 30, 1985.
[25] Point Cloud Library. Available: https://pointclouds.org/
[26] M. Vlaminck, H. Luong, and W. Philips, "Surface-based GICP," in 2018 15th Conference on Computer and Robot Vision (CRV), 2018: IEEE.
[27] H. Bay, T. Tuytelaars, and L. Van Gool, "Surf: Speeded up robust features," in European conference on computer vision, 2006: Springer.
[28] E. Rublee, V. Rabaud, K. Konolige, and G. Bradski, "ORB: An efficient alternative to SIFT or SURF," in 2011 International conference on computer vision, 2011: Ieee.
[29] L. Juan and O. J. I. J. o. I. P. Gwun, "A comparison of sift, pca-sift and surf," vol. 3, no. 4, 2009.
[30] E. Karami, S. Prasad, and M. J. a. p. a. Shehata, "Image matching using SIFT, SURF, BRIEF and ORB: performance comparison for distorted images," 2017.
[31] S. A. K. Tareen and Z. Saleem, "A comparative analysis of sift, surf, kaze, akaze, orb, and brisk," in 2018 International conference on computing, mathematics and engineering technologies (iCoMET), 2018: IEEE.
[32] D. G. Lowe, "Object recognition from local scale-invariant features," in Proceedings of the seventh IEEE international conference on computer vision, 1999, vol. 2: Ieee.
[33] D. G. J. I. j. o. c. v. Lowe, "Distinctive image features from scale-invariant keypoints," vol. 60, no. 2, 2004.
[34] M. Brown and D. G. J. I. j. o. c. v. Lowe, "Automatic panoramic image stitching using invariant features," vol. 74, no. 1, 2007.
[35] R. B. Rusu, N. Blodow, and M. Beetz, "Fast point feature histograms (FPFH) for 3D registration," in 2009 IEEE international conference on robotics and automation, 2009: IEEE.
[36] P. J. Besl and N. D. McKay, "Method for registration of 3-D shapes," in Sensor fusion IV: control paradigms and data structures, 1992, vol. 1611: International Society for Optics and Photonics.