点击次数：

论文类型：期刊论文

发表时间：2018-01-01

发表刊物：ARCHIVES OF METALLURGY AND MATERIALS

收录刊物：SCIE

卷号：63

期号：2

页面范围：673-680

ISSN号：1733-3490

关键字：longitudinal crack; visual detection; image processing; continuous casting

摘要：Based on the mould temperature measured by thermocouples during slab continuous casting, a difference of temperature thermograph is developed to detect slab cracks. In order to detect abnormal temperature region caused by longitudinal crack, the suspicious regions are extracted and divided by virtue of computer image processing algorithms, such as threshold segmentation, connected region judgement and boundary tracing. The abnormal regions are then determined and labeled with the eight connected component labeling algorithm. The boundary of abnormal region is also extracted to depict characteristics of longitudinal crack. Based on above researches, longitudinal crack with abnormal temperature region can be detected and is different from other abnormalities. Four samples of temperature drop arc picked up to compare with longitudinal crack on the abnormal region formation, length, width, shape, et al. The results show that the abnormal region caused by longitudinal crack has a linear and vertical shape. The height of abnormal region is more than the width obviously. The ratio of height to width is usually larger than that of other temperature drop regions. This method provides a visual and easy way to detect longitudinal crack and other abnormities. Meanwhile it has a positive meaning to the intelligent and visual mould monitoring system of continuous casting.