Research Of Key Technology Of Spectrum Detection Of Multi-Component Pollutants

Fourier transform infrared remote sensing is an effective way to detect the compositon and concentration of multi-component pollutants in the atmosphere far away, which has the advantage of high resolution, high signal to noise ratio, and high accuracy.Focus on the key technology of spectrum detection of multi-component pollutants, the work of this dissertation is listed as follows:(1) To tackle the problem of the noise and baseline drift which exist in the measured spectrum, a method has been proposed based on wavelet transform to remove both. This approach can be used to achieve baseline correction by setting the low frequent wavelet coefficients in the wavelet domain zero, and to remove noise by using the threshold to process the high frequent wavelet coefficients.(2) The experiment of wavelet denoising and baseline correction has been completed systematically. Optimal selection of the wavelet function, the decomposition scale and the threshold rules has been achieved through the three evaluation criterias: the normalized correlation coefficient (NCC), the root-mean-square error (RMSE) and the signal to noise ratio (SNR).(3) Experiment results have verified the approach of extracting principal component can remove the overlapping and redundant information of spectrum data quickly and compress the original spectrum without losing any important imformation of it. This method has resolved the problem of large calculation in the later qualitative and quantitative analysis effectively. So spectrum preprocessing can guarantee the accuracy of spectrum data and improve the speed of later calculation.(4) The hardware and software's designment of the spectrum data collection part of Fourier transform infrared spectrometer has been researched systematically.