A New Detection Method Applied on Steel Sectional Dimension Based on Image Processing

During the process of steel manufacturing, the sectional dimension would be detected to tell the quality of the steel. Nowadays the main detection method is by manual operation for machine. But the results of manual detection vary greatly according to the steel placement, operation process and subjectivity of the operators. Thus manual detection is wildly inaccurate. This paper introduces a new detection method applied on steel sectional dimension based on image processing. This method collects images of the steel cross-section, and converts images to binary ones. The edge feature points are extracted from the binarization images and steel sectional dimension would be detected by the maximum entropy method. The experimental results show that using this algorithm greatly improves the accuracy of detection and achieves satisfactory results.