双向反射分布函数结合Bi-LSTM网络求解壁面发射率

壁面光谱发射率求解是飞行器红外隐身的关键技术之一。首先设计了壁面反射光路和光源，通过光谱辐射计获取壁面反射的辐射亮度序列，为尽可能地消除外界干扰对于光谱发射率求解精度的影响，基于双向长短期记忆网络，设计了Bi-LSTM亮度回归网络模型，并对测试样本进行训练学习。基于BRDF的壁面发射率求解模型及基于Bi-LSTM网络的亮度回归模型求解壁面的发射率。计算结果显示，提出的基于双向反射分布函数的壁面发射率求解方法的相对误差为12.21%，满足工程测试需求。

Wall emissivity solved by bidirectional reflection distribution function combined with Bi-LSTM network

Wall spectral emissivity solution is one of the key techniques for infrared stealth of aircraft. Firstly, the wall reflected light path and light source were designed, and the brightness sequence of wall reflected radiation was obtained by spectral radiometer. In order to eliminate the influence of external interference on the solving accuracy of spectral emissivity as much as possible, Bi-LSTM brightness regression network model was designed based on bidirectional LSTM network, and the test samples were trained and learned. The wall emissivity solution model based on BRDF and the luminance regression model based on Bi-LSTM network were used to solve the wall emissivity. The calculation results show that the relative error of the proposed wall emissivity solution method based on BRDF is 12.21%, which meets the requirements of engineering test.