结合辐射传输和朗伯比尔定律, 提出一种基于有限脉冲响应数字滤波器和多元线性回归分析的空间外差干涉谱目标匹配方法, 通过对干涉图直接分析来提取出钾共振双线的特征以实现目标识别。本文采用传统有限脉冲响应数字滤波的处理方法来消除有效目标干涉频率信息以外的背景信息和干扰信号。通过对滤波前后的时域干涉数据进行多元线性回归处理获得最佳的时域滤波系数, 以对目标信号、背景信号的干涉图直接进行滤波处理, 最后通过干涉图匹配技术来识别钾共振双线信号。该算法无需测量背景光谱或背景干涉图, 亦无需对干涉图做傅里叶变换获取光谱, 在提高计算效率的同时提高空间外差光谱仪遥感数据的分析反演能力。

The algorithm of spatial heterodyne interferogram matching on finite impulse response and multiple linear regression analysis is proposed based on the radiation transfer model and Beer-Lambert law. For object identification, we extracted the characteristic of potassium double resonance from analyze interferogram directly. Traditional finite impulse response digital filter processing method was adopted to eliminate background information and other interfered information which except from interference frequency information of effective target. By multiple linear regression analysis processing interference data of time-domain of before and after filtering, the optimal coefficients of time-domain filtering had been obtained. So that, interferograms of target signals and background signals were filtered directly. Finally, the signal of potassium double resonance was identified by the interferogram matching. In this algorithm, it’s not required measuring the spectrum and interference pattern of background, neither Fourier transforming of interferogram to access spectrum, and improved efficiency and increased inverting capacity of spatial heterodyne spectroscopy analysis remote sensing data simultaneously.