论文部分内容阅读
实验测量到的谱带包络可认为是理论上的真实光谱被两种类型的实验畸变所干扰的结果:(a)随机噪声所引起的随机误差,(b)测量的实验条件和随后的数据处理中存在的固有的系统误差。没有随机噪声时,测量到的光谱(M(ν))可表示为真实光谱(E(ν))与仪器函数(G(ν))的褶积。实际上M(ν)是已知的,G(ν)可根据已知实验条件用公式表示出来。因此,原则上应可计算E(ν),它比实验的谱带轮廓有较锐的分离的谱带。但是,要取得计算结果是困难的,因为要求对M(ν)解褶积才得出E(ν)和G(ν)。下面举出由六个洛伦兹谱带重迭成的一组谱带的解褶积的例子。
The experimentally measured spectral envelope can be thought of as a result of theoretical real-world spectra being disturbed by two types of experimental distortions: (a) random errors due to random noise, (b) experimental conditions of the measurements and subsequent data Inherent systematic error in processing. Without random noise, the measured spectrum (M (ν)) can be expressed as the convolution of the true spectrum (E (ν)) and the instrument function (G (ν)). In fact M (ν) is known, and G (ν) can be formulated according to the known experimental conditions. Therefore, in principle, E (ν) should be calculated, which has a sharp separation of the spectrum from the experimental band profile. However, it is difficult to obtain the results because E (ν) and G (ν) are demanded to deconvolve M (ν). The following is an example of deconvolution of a set of bands that are overlapped by six Lorentzian bands.