摘要
基于数字图像处理技术来测量温度已成为高温测量的发展趋势.研究发现高温工件的灰度分布与温度存在单调关系,应用MATLAB对所测得45钢热处理加热时表面灰度数据进行处理,并采用BP神经网络和Levenberg-Marquarat优化算法建立了温度测量模型.该方法测量误差在4%以内,符合工业使用要求.
Measuring temperature based on digital image processing technology has become the develop- ment trend of high temperature measurement. Research results showed that a monotonous relationship ex- isted between temperature and the gray level distribution of the high temperature work-piece surface. In the present study, the gray data of 45# steel work-piece during the heat treatment process was processed with the MATLAB image processing technology, and a model of temperature measurement based on BP neural network and Levenberg-Marquarat optimization algorithm was established. The error of the meas- urement was below 4%, which meets the requirements of industry application.
出处
《西南大学学报（自然科学版）》
CAS
CSCD
北大核心
2013年第6期123-126,共4页
Journal of Southwest University(Natural Science Edition)
基金
重庆市科技攻关计划项目(CSTC
2011AC1016)
关键词
数字图像处理
神经网络
45钢
温度测量
digital image processing
neural network
45# steel
temperature measurement
