Study Of Medium Carbon Steel Slab The Prediction Methods Of Transverse Corner Crack

With the rapid development of iron and steel industry, steel billet production technology also has achieved great progress and improvement.It causes great influence on the quality of the production when the surface exist defects. At present,many domestic and foreign iron and steel enterprises still use the traditional manual detection methods.The way of testing is easily to human emotions, the testing will not be able to avoid the phenomenon such as leak and mistakenly identified, results in the decrease of slab quality, scrap rising, defective rate rising.Existing detection method based on image steel casting quality, which makes use of intelligent algorithm to recognize, casting surface quality identification will be affected by the limit of algorithm complexity,reduce the recognition efficiency.Therefore, avoiding the generation of surface defects of steel is of great important to improve the efficiency of production. Real-time detection for billet surface transverse crack prediction is also currently each big iron and steel enterprises facing with serious issue.By analyzing of the process of production environment, which is setting the infrared thermal imager in the exit of the mould, through analysis of thermal image acquisition to achieve on-line detection. Reducing the area noise effects on both end of the gray scale,improves the contrast of the image. Second using the average filtering method to make image smoothing. Further reducing the thermal image noise,Using different edge detection operator to segment of image corner crack, get high quality images. Selecting the Gauss Laplace on the slab surface transverse crack detection operator image recognition detection. Finally through regional marker method to calculate the feature information such as crack length, perimeter, area. The result shows that using infrared thermal image technology on surface quality of billet corner transverse crack detection system can be in steel slab continuous casting production of online real-time detection to obtain well application in the system.