印花織物是一種高附加價值的紡織品，具有豐富的色彩及多變化的圖案，是其它織物所不能及，惟雖然日常生活中使用的大都是彩色織物，但目前應用於織物的影像分析技術仍只對灰階織物。本研究之目的即在發展印花織物瑕疵自動檢測系統，以檢測印花織物數位影像並找尋其中瑕疵的影像，提高印花織物產品之品質。影像來源由掃描器將印花織物轉成數位影像資料，取得重覆圖樣色彩空間(Color Space)的紅綠藍(RGB)值，統計XY軸RGB累加平均值，比較XY軸上的RGB累加平均值可得到最小重覆區域並儲存。以最小重覆區域影像RGB累加平均值當作印花織物數位影像比較的依據，利用小波轉換去檢測輸入印花織物的數位影像是否有瑕疵，最後取得具瑕疵的影像時，統計瑕疵影像的RGB累加平均值與最小重覆區域的RGB累加平均值，經比對後系統會顯示有瑕疪區域的影像。最後比較最小重覆影像與瑕疪區域的影像可得到瑕疵影像部位，利用模糊分類器檢測瑕疪種類。由實驗結果顯示，此系統可分析出96.8%的印花織物瑕疵的種類，對印花織物自動檢驗具有很好成效。

Printed fabric is a high value-added textile that has rich colors and variable patterns. Although colored fabrics are commonly used in daily life, image analysis technology for fabrics is limited to gray-scale fabrics. This study aimed to develop an automatic defect detection system for printed fabrics that detects the digital images of printed fabrics and identifies the defected images, in order to enhance the quality of printed fabric products. The images are scanned from printed fabrics and converted to digital data. After obtaining the RGB values of repeated Color Space, the RGB accumulative average mean of XY axles are calculated and compared to obtain and save the minimum repeating zone. The RGB accumulative average mean of the minimum repeating zone is used as the basis of comparison for digital images of printed fabrics, and wavelet transform is used to detect the flaws in the digital images of printed fabrics. After the flaws are identified, the RGB accumulative average mean of the flawed images and minimum repeating zone are calculated and compared to indicate the images of the flawed zone. Lastly, the images of the minimum repeating zone and defect image are compared using fuzzy classifier to detect the type of defect. The experimental results showed that this system can analyze 96.8% of the detect types of printed fabrics, and has practical values in automatic detection of the quality of printed fabrics.

[1]呂勝昌，郭文貴，「影像傳輸在平版印花之應用黑圖繪製在現性之分析」，華岡紡織期刊，第五卷，第二期，第112-119頁，1998.
[2]P. Bernard, K. Vanderhaeghen, and P. Kiekens, “Color Measurement of Pile Fabrics”, Textile Research Journal, Vol. 69, No. 8, pp. 552-558, 1999.
[3]D. E. Akin, H. H. Epps, D. D. Archibald, and H. S. Shekhar Sharma, “Color Measurement of Flax Retted by Various Means”, Textile Research Journal, Vol. 70, No. 10, pp. 852-858, 2000.
[4]H. J. Lung, “Application and Development of Automatic Color Separating System for Print,” Journal of the China Textile Institute, Vol. 2, No. 4, pp. 24-31, 1992.
[5]H. J. Lung, “Development of Automatic Color Separating System for Print”, Journal of the China Textile Institute, Vol. 4, No. 2, pp. 111-115 (1994).
[6]C. Y. Shih, M. S. Hsu, and K. L. Hsu, “Color Pattern Analysis of Printed Fabric by Fuzzy Clustering Method”, Journal of the China Textile Institute, Vol. 21, No. 1, pp. 69-80, 2003.
[7]C. F. J. Kuo, C. Y. Shih, and J. Y. Lee, “Automatic Recognition of Fabric Weave Patterns Fuzzy C-Means Clustering Method”, Textile Research Journal, Vol. 74, No. 2,pp. 107-111, 2004.
[8]K. N. Plataniotis, and A. N. Venetsanopoulos, “Color Image Processing and Applications”, Springer, New York, pp. 239-246, 2000.
[9]L. A. Zadeh, “Outline of a new approach to the analysis of complex systems anddecision processes”, IEEE Transactions on Systems, Man and Cybernetics, Vol.SMC-3, No.1, pp. 28-44, 1973.
[10]B. Kosko, “Neural Networks and Fuzzy Systems”, Prentice -Hall, 1992.
[11]王進德，蕭大全，「類神經網路與模糊控制理論入門」，全華圖書，2007。
[12]T. D. Pham, and H.Yan, “Color image segmentation using fuzzy integral and mountain clustering”, Department of Electrical Engineerin 9, University of Sydney, Sydney, Australia, Fuzzy Sets and Systems Vol 107, Issue 2, pp. 121-130, 1999.
[13]S. W. Zucker, and D. Terzopoulos, “Finding Structure in Cooccurrence Matrices for Texture Analysis”, Proc, Int’l Conf. Color in Graphics and Image Processing, Vol.12, No.3, pp. 286-308, 1980.
[14]J. Russ, “The Image Processing Handbook”, second ed., CRC Press, 1995.
[15]H. C. Lin, and L. L. Wang, and S. N. Yang, “Extracting Periodicity of a Regular Texture Base on Autocorrelation Functions”, Pattern Recognition Letters, Vol. 18, pp. 433-443, 1997.
[16]C. Sanby, L. Norton-Wayne and R. Harwood, “The Automated Inspection of Lace Using Machine Vision”, Mechatronics, Vol. 5, No.2/3, pp. 215-231, 1995.
[17]M. Bradshaw, “The Application of Machine Vision to the Automated Inspection of Knitted Fabrics”, Mechatronics Vol. 5, No.3/3, pp. 233-243, 1995.
[18]W. Y. Wu, M. J. Wang, and C. M. Liu, “Automated inspection of printed circuit boards through machine vision”, Computers in Industry, vol. 28, pp. 103-111, 1996.
[19]H. Raafat, and S. Taboun, “An Integrated Robotics and Machine Vision System for Surface Flaw Detection and Classification”, Computers in Industrial Engineers, Vol.30, No1, pp. 27-40, 1996.
[20]P. G. Lemarie, and Y. Meyer, “Ondelettes et bases Hilbertiennes”, Rev. Mater. Ibero Am. 2, pp. 1–18, 1986.
[21]S. G. Mallat, “A theory for multiresolution signal decomposition: the wavelet representation”, IEEE Trans. Pattern Anal. Mach. Intell, pp. 674–693, 1989.
[22]A. Laine, and J. Fan, “Texture classification by wavelet packet signatures”, IEEE Trans. Pattern Anal. Mach. Intell, pp. 1186–1191, 1993.
[23]M. Unser, “Texture classification and segmentation using wavelet frames”, IEEE Trans. Image Process. pp. 1549–1560, 1995.
[24]Y. Chitre, and A.P. Dhawan, “M-band wavelet discrimination of natural textures”, Pattern Recogn. 32, pp. 773–789, 1999.
[25]A.L. Amet, A. Ertuzun, and A. Ercil, “Texture defect detection using subband domain co-occurrence matrices”, Image Anal. Interpretation, pp. 205–210, 1998.
[26]P.S. Windyga, “Fast impulsive noise removal”, IEEE Trans. On Image Processing, pp. 173-179, 2001.
[27]M.S. Nixon, and A.S. Aguado, “Feature Extraction and Image Processing”, Newnes, Oxford, pp. 89-95, 2002.