基于图像处理的钢坯定尺定重智能切割系统

来源期刊：中南大学学报(自然科学版)2002年第6期

论文作者：安辉耀 沈德耀

文章页码：643 - 647

关键词：图像处理;定尺定重切割;模式识别;自学习;视频信号

Key words：image processing; limited length and limited weight cutting; pattern recognition; self-learning; video signal

摘    要：针对目前国内切割工艺技术存在的缺陷及不足,提出了利用模式识别及图像处理技术,确定定尺长度,并以长度为基础,利用闭环反馈装置最终实现定重切割的方法.应用结果表明:系统利用摄像头远距离采集生产现场工艺参数,能代替操作人员在恶劣环境下工作,且操作和维护方便、可靠,加强了人员、设备的安全性;抗干扰能力强,判断准确;系统采用模式识别技术能很好地识别钢坯,因而对于外部来的非钢坯等产生的干扰信号能很好地识别,不会发生误差或者漏判等现象;精度高,将钢坯切割定尺误差由原来的80 mm减少到10~15 mm,定重相对误差小于0.5%,降低了生产成本,提高了成材率.

Abstract: Rough cutting section and excess error will cause more waste and increase cost. This paper studies pattern recognition and image processing, proposes a method to implement limited length and limited weight cutting with pattern recognition and image processing technique. The experiment results show that the system is satisfactory in operation,maintenance, cutting precision,processes levels and rates of finished products. Collecting work field′s data with cameras from long-distance, this method will not only reduce the possibility for people towork in the bad circumstance, but also operate more expediently and more reliably, making the equipment and the worker more safe. Strong anti-jamming can recognize the steel well and truly with these technique. It can reduce the length error from 80 mm to 10~15 mm, so it can reduce product cost, and increase success rate of the finished product.