论文部分内容阅读
干涉图滤波是干涉光谱成像仪光谱反演过程中的一个关键环节,常用的滤波方法主要是差分法和拟合法。差分法对背景噪声不能完全去除;拟合法则需要先验知识,而且在干涉数据两端拟合误差较大。经验模态分解(EMD)方法是近年来提出的一种新的用于线性和稳态谱分析信号处理方法,该方法提出后在很多领域得到广泛应用。将EMD方法应用到干涉图的滤波过程中,使得对背景噪声的提取更为合理,而且具有自适应性,避免了常用滤波方法的不足。利用实验室实际获取的数据进行分析,可以看出:EMD滤波后空间维的光谱相对均方根误差(RQE)均值为0.0068,精度最高;其次为拟合法,RQE均值为0.0073;最后为差分法,RQE均值为0.0079。
Interferogram filtering is a key step in the spectral inversion of interference spectrum imager. The commonly used filtering methods are mainly differential method and fitting method. The difference method can not remove the background noise completely; the fitting rule needs prior knowledge, and the fitting error is larger at the two ends of the interference data. Empirical Mode Decomposition (EMD) is a new signal processing method proposed in recent years for linear and steady-state spectral analysis. The proposed method is widely used in many fields. Applying the EMD method to the filtering process of the interferogram makes the extraction of the background noise more reasonable and adaptable and avoids the disadvantages of the commonly used filtering methods. Using the data obtained from the laboratory, we can see that the mean square root mean square error (RQE) of the space dimension after EMD filtering is 0.0068, the highest precision is obtained, followed by the fitting method, the RQE mean is 0.0073, and finally the difference method , RQE mean is 0.0079.