摘要
为了处理红外测温标定实验所得到的数据和建立温度灰度标定数学模型,介绍了ε-支持向量回归机基本原理及其在红外测温标定实验中的应用。在黑体温度从30~72℃变化过程中,采集22组实验样本,其中17组为训练样本,其余为预测样本。在数据处理时引进ε-支持向量回归机方法,且通过训练样本与预测样本在MATLAB下拟合出模型曲线。与传统的最小二乘法比较,ε-支持向量回归方法具有较高的精度,可以成为一种红外测温标定实验数据处理方法。
To process datas acquired from infrared temperature demarcating experiment and establish mathematical model between temperature and gray value, the basic principle of ε-SVR and Application of ε-SVR in infrared temperature demarcating experiment are introduced in this paper. In the process of the temperature of black body ranging from 30℃ to 72℃, 22 groups of samples are acquired, which include 17 groups of training samples and 5 groups of forecasting samples. The method of ε-SVR is introduced in data procession and the fitting curves are got through training samples and forecasting samples under the simulation of MATLAB. Compared with traditional method of least square, precision of this method is far higher. In conclusion, the method of ε-SVR can become a method of data processing to infrared temperature demarcating.
出处
《红外技术》
CSCD
北大核心
2009年第4期199-201,214,共4页
Infrared Technology
基金
质检公益性行业科研专项经费支持项目(项目编号:2007GYJ016)