热轧带钢在线测宽中的图像分割

Image Segmentation in On-line Width Measurement of Hot Strip

对带钢质量检测中图像处理和分析的过程进行了论述.结合热轧带钢现场图像,应用最终测量精度准则对图像处理算法进行评价.实验评价结果表明,动态双阈值的图像分割效果最佳,适合应用于带宽实时在线检测.

Some algorithms about image processing and analysis in strip quality inspection were discussed. With plenty of strip images being captured from a hot-rolled steel strip line, evaluation and comparison to these algorithms were done with the rule of UMA (Ultimate Measurement Accuracy), which showed that the dynamic threshold image segmentation was the best and could be applied to the on-line width measurement of hot strip.