THE SURFACE QUALITY OF CASTING DEFECT DETECTION SYSTEM

摘要

The surface defects of continuous casting for example the cracks, slag pit and double pouring often affect the quality of billet seriously. Using the surface defect detection system of billet which detects the defects by image analysis can improve the billet quality greatly. In this paper, we introduces the principle and structure of casting surface defect detection system and image processing steps and the method of extracting defect characteristics and classification. We selects the gray level co-occurrence matrix texture features and gray features and shape features, with them trained and classified by BP neural network. The effectiveness of the feature selection and classification method has been verified in the experiment.