臺灣博碩士論文加值系統

目前對於內螺紋常見的測量方式為使用GO-NO GO螺紋檢查牙規以人工方式進行檢測，扣件螺帽如符合允收標準，螺帽之半徑應對稱於軸心線，如扣件螺帽於製造中產生軸心線偏移問題，則會形成所謂的斜牙螺帽，斜牙螺帽之軸心線將與原始軸心線形成一夾角。

本研究以螺帽之內螺紋為檢測載具，應用機器學習針對螺帽斜牙進行研究，並利用五種模型評估指標進行評估，分別為準確度（Accuracy）、精確度（Precision）、召回率（Recall）、F1-Score以及科恩卡帕係數（Cohen’s kappa coefficient）。其中，精確度與召回率因其需考量使用環境，一般情形下希望兩者權重合理分配，為綜合考量而產生F1-Score，以及為避免單類樣本數占比過高影響準確度，產生了考慮偶發性高度一致的可能性，平衡類別之間正確性，把數量占比過高的某個類別消除的指標kappa。因前三項模型評估指標皆有偏頗及不適應性，後續討論將針對F1-Score與kappa。

對確認影像旋轉角度對模型偵測能力的影響進行模型評估，依據數據判斷訓練用影像旋轉360度要比旋轉正負180度的正確率高，而樣本數與影像雜訊對正確率的影響也是明顯的，Type3評估數據F1-Score僅有29.88%，Type4評估數據F1-Score則是98.70%，差距明顯。

Currently, the standard measurement method for the internal thread is to use a GO-NOT GO thread inspection gauge to detect manually. Suppose the fastener nut meets the acceptable standard. In that case, the radius of the nut should be symmetrical to the axis line. If there is an axis deviation problem during manufacturing, a so-called oblique thread nut will be formed, and the axis line of the oblique thread nut will form an included angle with the original axis line.

This study uses the internal thread of the nut as the detection vehicle, applies machine learning to study the oblique teeth of the nut, and uses five model evaluation indicators for evaluation, Accuracy, Precision, Recall, F1-Score, and Cohen's kappa coefficient. Among them, because Precision and Recall need to consider the use environment, the two weights should be reasonably distributed. The F1-Score is generated carefully to avoid the high proportion of single-class samples from affecting the accuracy. Consider the possibility of sporadic high consistency, balance the correctness between categories, and eliminate the index kappa of a specific type with an excessively high proportion. Since the first three model evaluation indicators are biased and inadaptable, the follow-up discussion will focus on F1-Score and kappa.

Carry out the model evaluation to confirm the impact of image rotation angle on the model's detection ability. Based on the data, it is judged that the correct rate of rotating the training image by 360 degrees is higher than that of rotating plus or minus 180 degrees, and the impact of the number of samples and image noise on the correct rate is also the same. Obviously, Type3 evaluation data F1-Score is only 29.88%, and Type4 evaluation data F1-Score is 98.70%; the gap is obvious.

摘要
Abstract
誌謝
目錄
表目錄
圖目錄
第一章 緒論
1.1 前言
1.2 研究動機
1.3 文獻回顧
1.3.1 文獻內容簡述
1.3.2 文獻回顧總結
1.4 論文架構
第二章 基礎理論
2.1 機器學習
2.2 ResNet
2.2.1 ResNet50
第三章 實驗設備與軟體
3.1 硬體設備介紹
3.1.1 內螺紋取像裝置
3.1.2 筆記型電腦設備
3.1.3 其他小工具
3.2 軟體工具介紹
3.2.1 Microsoft Visual Studio C#
3.2.2 EmguCV
3.2.3 ML.NET
第四章 研究方法
4.1 內螺紋影像擷取
4.2 內螺紋影像資料擴增
4.2.1 影像載入
4.2.2 旋轉參數設定
4.2.3 影像旋轉存檔
4.3 機器學習模型訓練與預測
4.3.1 單張影像檢測
4.3.2 大量影像檢測
4.4 模型評估指標
4.4.1 準確度（Accuracy）
4.4.2 精確度（Precision）與召回率（Recall）
4.4.3 F1-Score
4.4.4 科恩卡帕係數（Cohen’s kappa coefficient）
第五章 結果與討論
5.1 ML.NET 機器學習結果
5.2 機器學習模型評估
5.2.1 影像旋轉角度對模型偵測能力的影響
5.2.2 樣本數與影像雜訊對正確率的影響
5.3 模擬內螺紋斜牙缺陷檢測
第六章 結論與未來展望
6.1 結論
6.2 未來展望
參考文獻
附錄一 Type1 準確度曲線圖數據表
附錄二 Type2 準確度曲線圖數據表
附錄三 Type3 準確度曲線圖數據表
附錄四 Type4 準確度曲線圖數據表

[1] 黃一宸，中華民國110年，運用深度學習檢測內螺紋螺牙之光學檢測技術，碩士論文，正修科技大學 機電工程研究所。
[2] 丁凡琇，中華民國110年，應用深度學習於內螺紋瑕疵檢測之研究，碩士論文，國立高雄科技大學 機械工程系碩士班。
[3] 馬辰維，中華民國110年，整合深度學習與自動光學檢測在缺件分類之應用，碩士論文，國立高雄科技大學 機械工程系碩士班。
[4] Yann LeCun, Léon Bottou, Yoshua Bengio, and Patrick Haffner, 1998, “Gradient-Based Learning Applied to Document Recognition”,Proceedings of the IEEE ( Volume: 86, Issue: 11, November 1998) pp.2278-2324, IEEE.
[5] Alex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton, 2012"ImageNet Classification with Deep Convolutional Neural Networks", Proceedings of the 25th International Conference on Neural Information Processing Systems - Volume 1, pp.1097-1105.
[6] Karen Simonyan, Andrew Zisserman, 2015, “Very Deep Convolutional Networks for Large-Scale Image Recognition”, 3rd International Conference on Learning Representations (ICLR 2015), pp.1-14, Computational and Biological Learning Society.
[7] Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun, 2016, “Deep Residual Learning for Image Recognition”, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, Las Vegas, NV, USA.
[8] Kyu Beom Lee, Hyu Soung Shin, 2019, “An Application of a Deep Learning Algorithm for Automatic Detection of Unexpected Accidents Under Bad CCTV Monitoring Conditions in Tunnels”, International Conference on Deep Learning and Machine Learning in Emerging Applications (Deep-ML), IEEE, Istanbul, Turkey.
[9] Wang Wang, Guangze Wen, Zikun Zheng, 2022, “Design of Deep Learning Mixed Language Short Text Sentiment Classification System Based on CNN Algorithm”, 2nd International Conference on Mobile Networks and Wireless Communications (ICMNWC), IEEE, Tumkur, Karnataka, India.
[10] Xiaodong Zhai, Fei Qiao, 2020, “A Deep Learning Model with Adaptive Learning Rate for Fault Diagnosis”, 9th Data Driven Control and Learning Systems Conference (DDCLS), IEEE, Liuzhou, China.
[11] Zhongnan Zhao, 2022, “Research on Single-Character Image Classification of Tibetan Ancient Books Based on Deep Learning”, 3rd International Conference on Computer Vision, Image and Deep Learning & International Conference on Computer Engineering and Applications (CVIDL & ICCEA), IEEE, Changchun, China.
[12] Christian Hary, Satria Mandala, 2022, “Object Detection Analysis Study in Images Based on Deep Learning Algorithm”, International Conference on Data Science and Its Applications (ICoDSA), IEEE, Bandung, Indonesia.
[13] Supriya V. Mahadevkar, Bharti Khemani, Shruti Patil, Ketan Kotecha, Deepali R. Vora, Ajith Abraham, Lubna Abdelkareim Gabralla, "2022A Review on Machine Learning Styles in Computer Vision—Techniques and Future Directions”, IEEE Access ( Volume: 10), pp.107293-1073329.
[14] Bendjillali Ridha Ilyas, Beladgham Mohammed, Merit Khaled, Kamline Miloud, 2019, “Enhanced Face Recognition System Based on Deep CNN”, 6th International Conference on Image and Signal Processing and Their Applications (ISPA), IEEE,Mostaganem, Algeria.
[15] Tianpeng He, Wenzheng Li, Xijia Du, Huina Yang, Haoxi Cong, 2021, “Garbage Classification and Recognition System Based on Neural Network”, International Conference on Advanced Electrical Equipment and Reliable Operation (AEERO), IEEE, Beijing, China.
[16] Naoki Takisawa, Syunji Yazaki, Hiroaki Ishihata, 2020, “Distributed Deep Learning of ResNet50 and VGG16 with Pipeline Parallelism”, Eighth International Symposium on Computing and Networking Workshops (CANDARW), IEEE, Naha, Japan.
[17] Sheldon Mascarenhas, Mukul Agarwal, 2021, “A comparison between VGG16, VGG19 and ResNet50 Architecture Frameworks for Image Classification”, International Conference on Disruptive Technologies for Multi-Disciplinary Research and Applications (CENTCON), IEEE, Bengaluru, India.
[18] Peng Yang, Chaoyi Dong, Xiaoyi Zhao, Xiaoyan Chen, 2020, “The Surface Damage Identifications of Wind Turbine Blades Based on ResNet50 Algorithm”, 39th Chinese Control Conference (CCC), IEEE, Shenyang, China.
[19] Susana M. Vieira, Uzay Kaymak, João M. C. Sousa, 2010, “Cohen's kappa Coefficient as A Performance Measure For Feature Selection”, International Conference on Fuzzy Systems, IEEE, Barcelona, Spain.
[20] Juan Wang, Yongyi Yang, Bin Xia, 2019, “A Simplified Cohen’s Kappa for Use in Binary Classification Data Annotation Tasks”, IEEE Access (Volume:7), pp.164386-164397.
[21] 黃衛東, 陳碑青, 戴春勤，2012，一種螺絲斜牙檢出裝置，中華人民共和國國家知識產權局，N202539438U。
[22] 郭曉鋒, 趙曉宇，2010，一種螺絲內斜牙缺陷檢測系統及其檢測方法，中華人民共和國國家知識產權局，CN102042988A。
[23] Matthew D. Zeiler, Rob Fergus, 2014, “Visualizing and Understanding Convolutional Networks”, Computer Vision – ECCV 2014, pp.818–833, ECCV
[24] Christian Szegedy, Wei Liu, Yangqing Jia, Pierre Sermanet, Scott Reed, Dragomir Anguelov, Dumitru Erhan, Vincent Vanhoucke, Andrew Rabinovich, 2015, “Going Deeper with Convolutions”, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, Oston, MA, USA
[25] Md. Tanzil Shahriar, Yashna Islam, Md. Nur Amin, 2019, “Player Classification Technique Based on Performance for A Soccer Team Using Machine Learning Algorithms”, IEEE International Conference on Electrical, Computer and Communication Technologies (ICECCT), Coimbatore, India
[26] 余彥伯，中華民國111年，基於深度學習架構之空心鋁擠型模具設計研究，碩士論文，國立高雄科技大學 機械工程系碩士班。
[27] Limei Song, Xinyao Li, Yangang Yang, Xinjun Zhu, Qinghua Guo and Huaidong Yang, 2018, "Detection of Micro-Defects on Metal Screw Surfaces Based on Deep Convolutional Neural Networks", Sensors 2018, 18(11), 3709, DOI:10.3390/s18113709, MDPI
[28] Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun, 2016, “Identity Mappings in Deep Residual Networks”, Computer Vision – ECCV 2016 pp 630–645, ECCV.
[29] 巫建銘，中華民國111年，應用AI機器學習模擬方法分析結合力量控制裝置於六軸機械手臂之表面加工製程最佳化研究，碩士論文，國立中興大學，精密工程研究所。
[30] 游生全，中華民國108年，基於深度學習與雷射三角量測輔助檢測內螺紋之研究，碩士論文，國立交通大學 電控工程研究所。
[31] 聯信標準 螺紋標準類，1996，ISO一般用途公制螺紋－量測，ISO1502。
[32] 機器學習，維基百科，https://zh.wikipedia.org/zh-tw/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0
[33] 機器學習介紹，SAP官網，https://www.sap.com/taiwan/products/artificial-intelligence/what-is-machine-learning.html
[34] Microsoft C#官網，https://learn.microsoft.com/zh-tw/dotnet/csharp/
[35] Emgu CV官網，https://www.emgu.com/wiki/index.php/Main_Page
[36] Open CV 維基百科，https://zh.wikipedia.org/zh-tw/OpenCV
[37] Microsoft ML.NET 官網，https://learn.microsoft.com/zh-tw/dotnet/machine-learning/
[38] ILSVRC 歷屆的深度學習模型，https://chtseng.wordpress.com/2017/11/20/ilsvrc-%E6%AD%B7%E5%B1%86%E7%9A%84%E6%B7%B1%E5%BA%A6%E5%AD%B8%E7%BF%92%E6%A8%A1%E5%9E%8B/
[39] 李馨，2017，從零開始學Visual 2015 C#程式設計，博碩文化出版。