Automatic Inspection for Fabric

• Main Authors: Chih-Hao Lin, 林志豪
• Other Authors: Wen-Yen Wu
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/50352694126906892570

碩士 === 義守大學 === 工業工程與管理學系碩士班 === 94 === In recent years, the automation inspection is constantly proposed and be applyed. Previous researches demonstrate that the inspection works very well and sustain long. time. Thus, many enterprises begin to use automation machine system, and hope to replace the human inspections. This study use the optimal Gabor filter to construct the inspection method of fabric. In Gabor filter, the Gabor transform is composed by the concept of Gaussian function and exponential function, which include three parameters: bandwidth, frequency and orientation. In this paper, we perform maximum ratio method and real-coded genetic algorithm to determine the optimal parameter, and use the automatic inspection method to test the flaws of the fabric. In the detection of the flow, we first grab the image by CCD. The image we grabbed is not suitable to detect the fabric flaw without image pre-processing. After that, we use optimal Gabor to inspect fabric flaw. In order to show the result of detection, thresholding is used. In this paper, we use to select the threshold. Then compute aspect ratio, centroid and ratio of black spot to white one. Then we recognize the kind of the flaw by the way of the grey relational analysis This study experiment in five kinds flaw. The results of experiment show that the accuracy rate is 97.33%. The fake alarm rate is 0.67%. The recognition rate is 94.67%. This method can be used to detect fabric flaw effectively.