基于支持向量机的红外光谱发射率计算方法  

Infrared spectral emissivity calculating method based on support vector machine

作　　者：傅莉 刘丰瑞 王俊华 宫禹 FU Li;LIU Feng-rui;WANG Jin-hua;GONG Yu(School of Automation,Shenyang Aerospace University,Shenyang 110136,China;Aircraft Design and Research Institute of Shenyang,Shenyang 110135,China)

机构地区：[1]沈阳航空航天大学自动化学院,沈阳110136 [2]沈阳飞机设计研究所,沈阳110135

出　　处：《沈阳航空航天大学学报》2020年第2期43-48,共6页Journal of Shenyang Aerospace University

基　　金：国家自然科学基金(项目编号:61074090);辽宁省社会科学基金(项目编号:L12D2v070)。

摘　　要：光谱发射率在国防、医学等领域具有重要的研究意义和应用价值。建立一种基于支持向量机神经网络(SVM)的目标红外辐射亮度模型,利用这种模型对测试样本进行网络训练,建立3~5μm和8~12μm波段红外辐射亮度模型,进而计算不同波长下目标的光谱发射率。通过标准黑体的比对验证,所建SVM网络模型的光谱发射率与黑体实际的光谱发射率相比,最大相对误差小于1.1%,并将该方法应用于飞机隐身涂料钢板的光谱发射率计算。Spectral emissivity has important research significance and application value in the field of defense and medicine.This paper establishes a target infrared radiance model based on support vector machine(SVM)neural network.The model is used for network training of test samples,and infrared radiance models in the bands of 3 to 5μm and 8 to 12μm are established to calculate the spectral emissivity of the target at different wavelengths.Verification with standard blackbody is carried out.Compared with the actual spectral emissivity of blackbody,the maximum relative error of spectral emissivity of the developed SVM network model is less than 1.1%,and this method is applied to the calculation of stealth coating steel plate of aircraft.

关 键 词：支持向量机 光谱发射率 红外辐射亮度 黑体 飞机蒙皮 

分 类 号：TN219[电子电信—物理电子学]