基于图像识别的宽厚板轧机自动转钢方案

来源期刊：冶金自动化2010年第6期

论文作者：门全乐

文章页码：55 - 60

关键词：宽厚板； 数字图像识别； 自动转钢

Key words：wide heavy plate； digital image identification； automatic plate turning

摘    要：宽厚板轧机的自动转钢问题,是宽厚板轧制过程中的一个固有"瓶颈",由于目前还没有有效的解决方法,因此影响了轧钢过程控制的自动化水平。通过对现场生产情况的深入研究,提出了基于数字图像识别的解决方案。针对现场生产情况开发了一套图像处理算法,应用此套算法对生产现场拍摄的钢坯图像进行处理和识别,取得了令人满意的结果,表明基于图像识别的进坯控制方法在现实生产中是完全可行的,应用于生产实践能够进一步提高轧钢生产线的自动化水平。

Abstract: Automatic plate turning is a bottle-neck problem in fully automatic wide heavy plate rolling.However,there is no effective solution to the problem so far in practice.Based on the detailed analysis of rolling sequences and conditions,a digital image identification based scheme that facilitates the automatic plate turning is proposed in this paper.The corresponding image processing algorithm and application software are developed,and applied to processing of the plate images taken from practical rolling site.The result of the application is satisfactory,indicating that the proposed scheme is feasible.