- 學位論文
黃哲韻
Defect detection in stochastic textures using Fourier reconstruction
若您是本文的作者，可授權文章由華藝線上圖書館中協助推廣。
摘要
傳統紋路分析的應用大多將紋路影像透過傅立葉轉換後的頻譜來擷取與紋路相關的特徵指標，再利用不同分類器，例如貝氏機率、類神經網路等，來進行瑕疵分類或瑕疵檢測，而不同的紋路必須事先設計或選擇個別適用的指標組合。本研究提供一個應用於隨機性紋路表面的瑕疵檢測方法，利用影像轉換及影像還原的技術來去除具規律性之隨機紋路的特徵，而不需透過擷取紋路指標的方式來進行瑕疵檢測。 藉由傅立葉轉換，原始灰階影像中具規律性之隨機紋路的特徵表現於傅立葉頻譜上時，其頻率元素之功率會呈環狀分布，高功率強度的頻率元素會集中分布於中心點附近，一般而言功率強度會隨著環狀半徑的增大而遞減。本研究經由選取傅立葉頻譜上的一個最適當的半徑，將頻譜上中心點與半徑以外的頻率元素去除，再利用反傅立葉轉換以便將原始影像中具規律性的隨機紋路去除，而只保留瑕疵部份影像。由於本研究方法不需要事先儲存標準比對影像的資料，因此不受外在環境變動如光源、旋轉的影響。本研究經實驗樣本如砂紙、軟木塞、地毯、皮革及鑄件等紋路之驗證發現，對於具規律性之隨機紋路有極佳之檢測效果。
並列摘要
In this research we present a global approach for the automatic inspection of defects in randomly textured surfaces. The propose method does not rely on local features of textures. It is based on a global image restoration scheme using the Fourier transform. Since a stochastic texture has the surface of random irregularities, the spread of frequency components in the power spectrum space is isotropic and forms the shape approximate to a circle. A fine textured surface results in a large spread radius, whereas a coarse one yields a smaller spread radius. The power magnitude is generally decreased as the frequency component is away from the circular center in the spectrum space. By finding an adequate radius in the spectrum space, and setting the frequency components outside the selected circle to zero, we can remove the periodic, repetitive patterns of any stochastic texture using the inverse Fourier transform. In the restored image, the homogeneous region in the textured image will have an approximately uniform gray level, and the defective region will be distinctly preserved. A statistic process control scheme is therefore used to set up the control limits for discriminating between defects and homogeneous patterns. The experiments on a variety of stochastic textures including sandpaper, castings , leather, and carpets have shown the effectiveness of the proposed method.
參考文獻
2.謝志雲，「應用機器視覺於方向性紋路之表面瑕疵檢測」，碩士論文，私立元智大學工業工程研究所，中壢巿，1998年。
3.Amet, A. L., A. Ertuzun and A. Ercil, “Texture defect detection using subband domain co-occurrence matrices,” Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation, pp. 205-210, 1998.
5.Bichebois, P. and P. Mathery, “Analysis of defect to yield correlation on memories: method, algorithms and limits,” Proceedings of the International Symposium on Defect and Fault Tolerance in VLSI Systems, pp. 44-52, 1997.
6.Brigham, E. O., The Fast Fourier Transform and Its Applications, Prentice-Hall, Englewood Cliffs, 1994.
7.Bracewell, R. N., The Fourier Transform and Its Applications, Mcgraw-Hill, New York, 1978.