刀具狀態為影響金屬機械加工成品精度與良率的重要直接因素之一，其中刀具磨耗為加工中狀態不斷改變之因素，易造成加工精度下降與不良品非預期產生。因此在磨耗程度無法被準確評估的情況之下，汰換刀具成為業者維持加工成品品質的策略之一，但也同時造成生產成本的上升。有鑑於端銑刀為機加工廠所常用於銑銷加工工序的一種刀具，因此本研究以端銑刀做為研究對象。
一般業界對於刀具磨耗的評估方式受個人主觀意識、加工工序、工件材料等因素影響，缺乏客觀量化標準。故本研究期能從科學定義的磨耗樣態發展出量化的分類標準，取代原分級方式。本研究已先前建立的背照式光學系統為基礎，建立一套自動化光學檢測系統並優化其演算法並提取刀具磨耗資訊，再以刀具磨耗資訊進行分析以辨識出刀具上所存在不同的磨耗樣態，以提供機加工業決策者更多資訊，包含磨耗樣態與磨耗數值等決策指標，幫助優化機加工業的整體效能。

Tool condition is having one of the important role in affecting the precision and yield of product. Tool wear is a factor that continuously changing in the processing period, which can easily cause the decrease of machining accuracy and unexpected production of scrap products. In the case where the wear level cannot be accurately evaluated, tool replacement has become one of the strategies for the industry to maintain the quality of the finished product, but it has also caused an increase in the cost of production. Given that the end mill is a milling tool that is widely used in milling process at manufacturing industry. This research takes the end mill as the research subject.
By industry, the assessment of tool wear is affected by subjective consciousness, processing method, materials of the work-piece, and it lacks objective quantitative standards. Therefore, this research aims to develop a quantitative classification standard from the scientific definition of wear type to replace the industrial grading method. This research based on the previously built inspection system, rebuild an automatic optical inspection system, modify the reconstruction algorithm also generate the wear information. By analyzing the wear information, it could recognize existing wear type on the end mill. Users in manufacturing industry could make the decision based on the wear information, including wear type and the index of wear, and optimized the performance of the manufacturing.

[1] E. P. DeGarmo, J. T. Black, R. A. Kohser, and B. E. Klamecki, Materials and process in manufacturing. Prentice Hall Upper Saddle River, 1997.
[2] Tool life testing in milling, part 2: end milling.
[3] S. Ravikumar, K. Ramachandran, and V. J. E. s. w. a. Sugumaran, "Machine learning approach for automated visual inspection of machine components," vol. 38, no. 4, pp. 3260-3266, 2011.
[4] J. Su, C. Huang, and Y. J. J. o. M. P. T. Tarng, "An automated flank wear measurement of microdrills using machine vision," vol. 180, no. 1-3, pp. 328-335, 2006.
[5] W. Wang, Y. Wong, and G. S. J. C. i. I. Hong, "Flank wear measurement by successive image analysis," vol. 56, no. 8-9, pp. 816-830, 2005.
[6] W. Wang, G. Hong, Y. J. I. J. o. M. T. Wong, and Manufacture, "Flank wear measurement by a threshold independent method with sub-pixel accuracy," vol. 46, no. 2, pp. 199-207, 2006.
[7] M. T. García-Ordás, E. Alegre, V. González-Castro, and R. J. T. I. J. o. A. M. T. Alaiz-Rodríguez, "A computer vision approach to analyze and classify tool wear level in milling processes using shape descriptors and machine learning techniques," vol. 90, no. 5-8, pp. 1947-1961, 2017.
[8] M. T. García-Ordás, E. Alegre-Gutiérrez, V. González-Castro, and R. J. I. J. o. P. R. Alaiz-Rodríguez, "Combining shape and contour features to improve tool wear monitoring in milling processes," vol. 56, no. 11, pp. 3901-3913, 2018.
[9] X. Wu, Y. Liu, X. Zhou, and A. J. S. Mou, "Automatic Identification of Tool Wear Based on Convolutional Neural Network in Face Milling Process," vol. 19, no. 18, p. 3817, 2019.
[10] D. M. D’Addona, A. S. Ullah, and D. J. J. o. I. M. Matarazzo, "Tool-wear prediction and pattern-recognition using artificial neural network and DNA-based computing," vol. 28, no. 6, pp. 1285-1301, 2017.
[11] 王鴻鈞, H.-C. Wang, 葉哲良, 駱遠, J. A. Yeh, and Y. Luo, 應用三維機械視覺於端銑刀磨耗檢測系統. 新竹市: 國立清華大學, 2018, p. 92面.
[12] 莊于萱, 葉哲良, 駱遠, J. A. Yeh, and Y. Luo, 端銑刀磨耗定量檢測系統. 國立清華大學動力機械工程研究所, 民國108年7月.