A study on approaches to classification of surface defects on cold rolled strips

Steel industry customers demand consistently high quality of steel strips. Automatic on-line surface inspection systems have been applied for monitoring as well as enhancing strip surface quality. One of the important issues in the application of this type of equipment is the performance of the on-line defect classifier. In this work rule-based classification table methods and neural network approaches are examined. The enhanced classification tables which is newly proposed here and neural networks have shown very excellent performance as the most promising classifiers for classification of surface defects on celled rolled strips.