Abstract: In order to solve the accuracy for measured results,the intuitiveness、intelligence and non-contact for the measured process for melting temperature of continuous casting mold powder.Based on machine vision technology,a system for melting temperature of continuous casting mold powder is designed and realized.The system establishes a reasonable model of the visual system to eliminate the interference of high temperature scene light,selects a high-exposure industrial camera to image the sample clearly,uses image analysis algorithm to extract effective information from the image,and tracks and recognizes the sample in real time through image tracking technology.The field test shows that the system for melting temperature of continuous casting mold powder based on machine vision can complete the measurement for melting temperature accurately,and the repeatability error is within ±20 ℃,which meets the requirements of YB/T 186 for melting temperature of continuous casting mold powder.

Key words: machine vision, test system, continuous casting mold powder, Image recognition, Steelmaking

摘要： 为了解决保护渣熔化温度测量结果的准确性不足,测量过程的直观性欠缺、智能化程度低及非接触测量需求难以满足的问题。基于机器视觉技术,设计了一款保护渣熔化温度测定系统;系统通过建立合理的视觉系统模型来排除高温场景光的干扰,选用高曝光度工业相机对试样进行清晰成像,利用图像分析算法对被测试样的图像进行有效信息提取,通过图像跟踪技术,对试样进行实时跟踪与识别;现场试验表明:基于机器视觉的保护渣熔化温度测定系统能够实时准确地完成保护渣熔化温度的测量,重复性误差不超过±20 ℃,满足YB/T 186对保护渣温度检测的要求。

关键词: 机器视觉, 测试系统, 保护渣, 图像识别, 钢铁冶炼