The industrial on-line defect detection system based on machine vision has good characteristics such as high detection speed, high precision, non-contact. In the field of quality inspection of industrial products, industrial on-line defect detection systems have been widely used. In recent years, vision based surface defect detection system has become a hot field of research and application for industrial visual inspection. Although many scholars and research bodies proposed various vision based surface defect detection systems for specific objects, many difficult problems persist in the real progress of applications. This thesis is focused on three key techniques in the fields of surface defect detection system based on machine vision: system design, defect detection algorithm, and control and decision method. The main work can be summarized as follows: 1) In the current available cotton impurities detection system, both the equipments and the detection procedures seem complex and un-easy to use. We propose a vision based cotton contamination detection system based on gray-scale camera. A cotton impurities detection algorithm for gray-scale images is also proposed. Gray Level Co-occurrence Matrix (GLCM) algorithm is adopted to detect the sharp contrast objects. Clustering algorithm is used to extract the background. An edge tracing algorithm based on objects neighborhood points is designed to remove edges of fake defects. Experiments show that the proposed method is easy to use and efficient. 2) To detect cotton impurity in industrial environment with uneven illumination, an impurity detection algorithm based on Gabor filter is proposed. In the algorithm, the image is divided into multiple zones using Otsu's threshold method and morphological filter, and the texture features for foreground and background are then extracted using Gabor filter. An adaptive threshold segmentation method is designed and applied to the Gabor filter output, and then the impurities in the image are detected by morphological filter and connected-zone analysis.Experiments results show that the proposed algorithm is efficient to remove the undesired interference caused by light source fluctuations, and is capable of accurately detecting common impurities in cotton. 3) To measure the fluid velocity in the pneumatic conveying pipeline for raw cotton, A soft-measuring method based on neural network is proposed. The teacher signal of the soft-measuring model is the differenc...
修改评论