臺灣博碩士論文加值系統

渦輪葉片是航空器常見的關鍵設備，檢測渦輪葉片的健康狀態對確保裝置運行和生產效率至關重要。傳統的渦輪葉片檢測方法主要基於視覺和經驗，對操作人員的要求較高，容易出現主觀偏差和誤判。
本研究旨在利用影像處理技術輔助航空器引擎渦輪葉片的內視鏡檢測，以協助維修人員判斷損傷狀況。當維修員發現損傷位置時，將其記錄並截圖保存，再經由影像輸出至系統進行處理，利用Median Filter與Gaussian Filter對影像進行前處理以分辨葉片位置，將處理後的影像用Otsu Method與Adaptive Threshold Method進行影像分割，然後利用Hough Transform和Convex Hull提取特徵資訊，並建立數據庫輔助判斷損傷情況。實驗結果顯示，針對Broken和Crack類別的分類性能有較高的準確率，分別達到了86%和91%，而對於Normal類別的分類效果也有了明顯改善，達到了79%的準確率，雖然對於Erosion類別的分辨率仍然有限，但是利用GLCM的特性成功地將Erosion的準確率提升至78%。本研究成功開發了一個影像輔助的渦輪葉片內視鏡檢測系統，能夠提高檢測效率和準確性，同時減少了操作人員的操作難度和誤判機率，為渦輪葉片的健康狀態監控提供了新的解決方案。

Turbine blades play a critical role in aircraft systems, and monitoring their health is essential for ensuring operational efficiency. Traditional inspection methods heavily rely on visual examination and operator experience, leading to subjective biases and misjudgments.
This study improves turbine blade inspection using image processing techniques. An image-assisted borescope system aids maintenance personnel in assessing blade damage by recording and processing screenshots of damage locations. Pre-processing involves Median Filter and Gaussian Filter, while image segmentation utilizes Otsu Method and Adaptive Threshold Method. Feature extraction techniques like Hough Transform and Convex Hull establish a damage classification database.
Experimental results demonstrate improved accuracy in classifying the Broken and Crack categories, achieving 86% and 91% accuracies, respectively. The accuracy for the Normal category also significantly improved to 79%. Although the resolution for the Erosion category remains limited, the use of GLCM successfully enhances the accuracy to 78%.
This research successfully develops a borescope image inspection system for turbine blades, which enhances detection efficiency and accuracy while reducing operator complexity and misjudgments. It provides a new solution for monitoring the health status of blades.

摘要………i
Abstract………ii
誌謝………iii
目錄………iv
表目錄………vi
圖目錄………vii
第1章 緒論………1
1.1 研究背景與動機………1
1.2 研究目的………1
1.3 文獻回顧………2
1.4 論文概述………4
第2章 航空發動機簡介………5
2.1 渦輪螺旋槳發動機………5
2.2 葉片檢測方法與葉片分類準則………6
第3章 數位影像處理………9
3.1 影像前處理………10
3.1.1 Gaussian Filter………11
3.1.2 Median Filter………11
3.1.3 Edge Detection………12
3.1.4 Morphology………12
3.2 影像分割………14
3.2.1 Otsu Method………15
3.2.2 Adaptive Threshold Method………16
3.2.3 Clustering………17
3.3 表示與描述………18
3.3.1 Hough Transform………19
3.3.2 Convex Hull………20
3.3.3 Connected Component………21
3.3.4 Gray-Level Co-occurrence Matrix………22
第4章 研究內容與方法………23
4.1 實驗設備………25
4.2 數位影像處理步驟………27
4.2.1 影像切割………27
4.2.2 影像前處理………27
4.2.3 影像二值化………30
4.2.4 Crack損傷處理………33
4.2.5 Broken損傷處理………36
4.2.6 Erosion損傷處理………39
第5章 結果與討論………40
5.1 檢測方法之參數說明………41
5.1.1 Crack缺陷結果………41
5.1.2 Broken缺陷結果………48
5.1.3 Erosion缺陷結果………51
5.2 失敗方法之參數說明………54
5.3 檢測結果………56
第6章 結論與未來展望………57
參考文獻………58
Extended Abstract………60

[1]李益誠, "影像處理於渦扇發動機轉子葉片損傷之檢測," 碩士, 機械工程學系在職專班, 國立中央大學, 桃園縣, 2020.
[2]高建勇, "渦輪葉片瑕疵之影像自動化檢測系統研發," 碩士, 應用力學研究所, 國立臺灣大學, 台北市, 2000.
[3]王聖儒, "渦輪葉片的智慧型磁粒檢測方法," 碩士, 機械工程學系, 國立中山大學, 高雄市, 1999.
[4]H. Shang, C. Sun, J. Liu, X. Chen, and R. Yan, "Deep learning-based borescope image processing for aero-engine blade in-situ damage detection," Aerospace Science and Technology, vol. 123, p. 107473, 2022/04/01/ 2022.
[5]D. Zhang, N. Zeng, and L. Lin, "Detection of Blades Damages in Aero Engine," in 2020 Chinese Automation Congress (CAC), 2020, pp. 6129-6134.
[6]內政部空中勤務總隊, "渦輪螺旋槳發動機和渦輪風扇發動機之差異探討," ed, 1998.
[7]星橙天文, "飛機發動機，只知道噴氣式？原來種類有這麼多," ed, 2018.
[8]Emoscopes, "渦輪螺旋槳發動機的運行示意圖," M0tty, Ed., ed, 2009.
[9]陳永增．鄧惠源, 非破壞檢測(修訂版). 全華圖書股份有限公司.
[10]微視光電, "工業內窺鏡在航空航天領域是飛機例行檢查工具之一," ed, 2018.
[11]L. Marie-Victerin, Quebec, PT6A6 SERIES AND PT6A-20 SERIES ENGINES COMBINED MAINTENANCE MANUAL 下冊. 4 JULY 1973.
[12]A. McAndrew, 數位影像處理 A COMPUTATIONAL INTRODUCTION TO DIGITAL IMAGE PROCESSING 2/E. 2020.
[13]K. Bachman, Implementation of an Autofocusing Algorithm in a Light Sheet Micro-scope for Imaging Thick 3D Samples. 2016.
[14]R. A. Jarvis, "On the identification of the convex hull of a finite set of points in the plane," Information Processing Letters, vol. 2, no. 1, pp. 18-21, 1973/03/01/ 1973.
[15]K. R. Anderson, "A reevaluation of an efficient algorithm for determining the convex hull of a finite planar set," Information Processing Letters, vol. 7, no. 1, pp. 53-55, 1978/01/01/ 1978.
[16]R. M. Haralick, K. Shanmugam, and I. Dinstein, "Textural Features for Image Classification," IEEE Transactions on Systems, Man, and Cybernetics, vol. SMC-3, no. 6, pp. 610-621, 1973.
[17]D. Bradley and G. Roth, "Adaptive Thresholding using the Integral Image," J. Graphics Tools, vol. 12, pp. 13-21, 01/01 2007.
[18]L. He, X. Ren, Q. Gao, X. Zhao, B. Yao, and Y. Chao, "The Connected-Component Labeling Problem: A Review of State-of-the-Art Algorithms," Pattern Recognition, vol. 70, 04/01 2017.
[19]L. Chandrasekar and G. Durga, "Implementation of Hough Transform for image processing applications," in 2014 International Conference on Communication and Signal Processing, 2014, pp. 843-847.
[20]M. Oge, "Morphological Image Processing," in Practical Image and Video Processing Using MATLAB: IEEE, 2011, pp. 299-334.
[21]S. Kheirinejad, S. M. H. Hasheminejad, and N. Riahi, Max-Min Ant Colony Optimization Method for Edge Detection Exploiting a New Heuristic Information Function. 2018, pp. 12-15.
[22]X. Zheng and N. Liu, "Color recognition of clothes based on k-means and mean shift," in 2012 IEEE International Conference on Intelligent Control, Automatic Detection and High-End Equipment, 2012, pp. 49-53.