刀具磨耗在機械加工業一直是很重要的問題，刀具的品質以及磨耗程度都會直接影響成品的表面粗糙度以及尺寸是否符合要求。隨著智慧機械的普及，加工機不僅限於加工廠之內，甚至非機械專業背景的人員操作加工機的人口也越來越多，然而大多數的人並不清楚刀具磨耗的重要性以及汰換刀具的時機，而使用磨耗情況嚴重的刀具加工只會浪費製作成本與時間。
而現有的刀具磨耗檢測方式大多是採用線上的方式偵測，需要在加工機裡安裝許多感測器以量測即時狀態，並以間接的方式判斷刀具磨耗的狀態為何。這種方式不僅容易受到外在因素影響評估結果，其方法的成本過於昂貴，也大幅降低非機械專業背景的人使用的意願。
本研究提出一種基於影像處理的低成本線外刀具磨耗檢測方法，包括硬體設計與軟體程式撰寫，觀測的刀具為直徑為2 mm 的兩刃球頭銑刀。硬體方面，本研究設計並製作一個刀具磨耗檢測裝置，透過微型電腦控制攝影機與步進馬達旋轉並拍攝刀具側面的影像，確保拍攝到刀具切削處輪廓的最大值。軟體方面，本研究使用開源軟體Python撰寫影像處理以量測刀具切削處之尺寸大小。透過灰階化與裁切影像能降低處理高解析度影像的時間，接著二值化取出刀具輪廓，最後抓取刀具切削處邊緣並計算尺寸的變化。我們將使用本研究所開發的量測裝置所得到的量測結果與二次元影像儀的量測結果比較，差異約在 ± 5 µm以內。實驗設計方面，本研究量測三把刀具的三種磨耗情況，最後量測出270筆數據進行分析。本研究提出之刀具磨耗檢測方式可以量測刀具切削處磨耗前後的差異，其量測解析度為4.56 µm。並且不論刀具磨耗情況為何，誤差範圍皆可控制在± 5 µm以內。在重覆性(Repeatability)測試方面，在信心水準95%的情況下，量測結果之準確率為 ± 8 µm，遠小於於新刀與汰換刀的尺寸變化至少40 µm相比，故本研究提出的檢測方法應足以用來分辨刀具磨耗情況。

Tool wear is an important problem in the machining industry as it directly affects the finished products' surface roughness and dimensional accuracy. With the increase of smart machinery, machining equipment is no longer limited to manufacturing facilities, and even individuals without a mechanical background are increasingly operating machining machines. However, most people are unaware of the importance of tool wear and the appropriate timing for tool replacement, leading to wasted production costs and time when using heavily worn tools for machining.
Existing methods for tool wear detection mostly rely on online monitoring, which requires the installation of numerous sensors in the machining machine to measure real-time conditions and indirectly determine the state of tool wear. This approach is susceptible to external factors that can affect the evaluation results and incur high costs, significantly reducing the willingness of non-mechanical professionals to use such methods.
This study proposes a low-cost offline tool wear detection method based on image processing, including hardware design and software programming. The practical tool is a 2 mm diameter two-flute ball-end milling tool. In terms of hardware, a tool wear detection device is designed and fabricated. It utilizes a microcomputer to control a camera and a stepper motor to rotate and capture the tool's side-view images, ensuring the cutting edge's maximum contour is captured. In terms of software, image processing is performed using the open-source software Python to measure the dimensions of the tool's cutting-edge width. Grayscale conversion and image cropping reduce the processing time for high-resolution images. Subsequently, binarization is applied to extract the tool contour, capture the tool's cutting-edge width and calculate the dimensional changes.
The measurement results obtained from the developed device are compared with those from a 2D imaging measurement instrument, showing differences within ± 5 µm. In the experimental design, three different tool wear conditions are measured, resulting in a total of 270 data points for analysis. The proposed tool wear detection method effectively measures the dimensional differences in the cutting region before and after wear, with a measurement resolution of 4.56 µm. Furthermore, the measurement error can be controlled within ± 5 µm regardless of the tool wear level. Regarding repeatability testing, at a 95% confidence level, the measurement results exhibit an accuracy of ±8 µm, significantly smaller than the minimum dimensional change of 40 µm between new and worn-out tools. Thus, the proposed detection method is suitable for discerning tool wear conditions.

[1] Y. Dai and K. Zhu, "A machine vision system for micro-milling tool condition monitoring," Precision Engineering, vol. 52, pp. 183-191, 2018.
[2] U. Ullah, F. A. Bhatti, A. R. Maud, M. I. Asim, K. Khurshid, and M. Maqsood, "IoT-enabled computer vision-based parts inspection system for SME 4.0," Microprocessors and Microsystems, vol. 87, pp. 104354, 2021. doi:https://doi.org/10.1016/j.micpro.2021.104354.
[3] W. Watanabe, R. Maruyama, H. Arimoto, and Y. Tamada, "Low-cost multi-modal microscope using Raspberry Pi," Optik, vol. 212, pp. 164713, 2020.
[4] M. C. Gomes, L. C. Brito, M. B. da Silva, and M. A. V. Duarte, "Tool wear monitoring in micromilling using Support Vector Machine with vibration and sound sensors," Precision Engineering, vol. 67, pp. 137-151, 2021.
[5] R. G. Lins, B. Guerreiro, P. R. M. de Araujo, and R. Schmitt, "In-process tool wear measurement system based on image analysis for CNC drilling machines," IEEE Transactions on Instrumentation and Measurement, vol. 69, no. 8, pp. 5579-5588, 2019.
[6] J. Jurkovic, M. Korosec, and J. Kopac, “New approach in tool wear measuring technique using CCD vision system,” International Journal of Machine Tools and Manufacture, vol. 45, no. 9, pp. 1023-1030, 2005.
[7] M. Castejón, E. Alegre, J. Barreiro, and L. K. Hernández, “On-line tool wear monitoring using geometric descriptors from digital images,” International Journal of Machine Tools and Manufacture, vol. 47, no. 12, pp. 1847-1853, 2007.
[8] X. Wang, W. Wang, Y. Huang, N. Nguyen, and K. Krishnakumar, “Design of neural network-based estimator for tool wear modeling in hard turning,” Journal of Intelligent Manufacturing, vol. 19, no. 4, pp. 383-396, 2008.
[9] K. Niranjan Prasad and B. Ramamoorthy, “Tool wear evaluation by stereo vision and prediction by artificial neural network,” Journal of Materials Processing Technology, vol. 112, no. 1, pp. 43-52, 2001.
[10] T. Bergs, C. Holst, P. Gupta, T. Augspurger, Digital image processing with deep learning for automated cutting tool wear detection, Procedia Manuf. 48, pp. 947–958, 2020.
[11] H. Liu, Z. Liu, W. Jia, D. Zhang, Q. Wang, and J. Tan, "Tool wear estimation using a CNN-transformer model with semi-supervised learning," Measurement Science and Technology, vol. 32, no. 12, pp. 125010, 2021.
[12] M. P. Kumar, S. Dutta, and N. Murmu, "Tool wear classification based on machined surface images using convolution neural networks," Sādhanā, vol. 46, no. 3, pp. 1-12, 2021.
[13] J. Ma, D. Luo, X. Liao, Z. Zhang, Y. Huang, and J. Lu, "Tool wear mechanism and prediction in milling TC18 titanium alloy using deep learning," Measurement, vol. 173, pp. 108554, 2021.
[14] Z. Li, X. Liu, A. Incecik, M. K. Gupta, G. M. Królczyk, and P. Gardoni, "A novel ensemble deep learning model for cutting tool wear monitoring using audio sensors," Journal of Manufacturing Processes, vol. 79, pp. 233-249, 2022.
[15] H. Oo, W. Wang, and Z. Liu, "Tool wear monitoring system in belt grinding based on image-processing techniques," The International Journal of Advanced Manufacturing Technology, vol. 111, no. 7, pp. 2215-2229, 2020.
[16] C. Zhang and J. Zhang, "On-line tool wear measurement for ball-end milling cutter based on machine vision," Computers in industry, vol. 64, no. 6, pp. 708-719, 2013.
[17] A. Agarwal, N. Potthoff, A. M. Shah, L. Mears, and P. Wiederkehr, "Analyzing the evolution of tool wear area in trochoidal milling of Inconel 718 using image processing methodology," Manufacturing Letters, vol. 33, pp. 373-379, 2022.
[18] K. Zhu and X. Yu, “The monitoring of micro milling tool wear conditions by wear area estimation,” Mechanical Systems and Signal Processing, vol. 93, pp. 80-91, 2017.
[19] L. Li and Q. An, “An in-depth study of tool wear monitoring technique based on image segmentation and texture analysis,” Measurement, vol. 79, pp. 44-52, 2016.
[20] M. Szydłowski, B. Powałka, M. Matuszak, and P. Kochmański, “Machine vision micro-milling tool wear inspection by image reconstruction and light reflectance,” Precision Engineering, vol. 44, pp. 236-244, 2016.