Fractal analysis of complicated textures

• Main Authors: Jiun-Jian Liaw, 廖俊鑑
• Other Authors: Shen Chou
• Format: Others
• Language: zh-TW
• Published: 2001
• Online Access: http://ndltd.ncl.edu.tw/handle/79180603436422342632

碩士 === 國立臺灣科技大學 === 纖維及高分子工程系 === 89 === In application of automatic quality control can be enhanced efficiency and decreased cost by computer interchanges human work. Digital images that have lots of advantages over analog images can be perfectly stored, transmitted and reprocessed. In the automatic inspection of texture surface defects, the most important technology is to segment the defects from the background. However, for complicated textures, the defect segmentation is still a challenging problem. Irregular pattern of nature can be described by fractal geometry. Lately, the fractal analysis method has been applied to complicated textures inspection. In this paper, we apply the Fractional Brownian Motion (FBM) model and Fourier-domain Maximum Likelihood Estimator (FDMLE) to compute fractal parameter for make an inspection of surface defects of textile fabrics. At last, the experiments was presented for prove the method works. It is applied in digital images. From the experimental results, we obtain good results of defect segmentation, and find the method’s performance is invariant under geometric transformation.