【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.更多还原

【Abstract】 Measuring temperature based on digital image processing technology has become the development trend of high temperature measurement.Research results showed that a monotonous relationship existed 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 measurement was below 4%,which meets the requirements of industry application.更多还原