Proceedings PaperDetermination of plume temperature distribution based on the ratios of the radiation intensities of multiple CO2 lines
|Format||Member Price||Non-Member Price|
New inversion scheme for gas temperature distribution retrieval utilized CO2 spectrum between 2350 cm-1 and 2400 cm-1 is proposed. Inversion model is build base on neural networks. Considered spectral remote sensing method is commonly used for industrial and environmental monitoring. It is a passive single-ended sensor technique in which radiation intensity emerging from a studied object is analyzed. Quantitative investigation of heated gas radiation emission to determine temperature and gas mixture by infrared spectroscopy requires two components apart from optical radiation sensor. First appropriate spectral database and second efficient inversion techniques. In this study calculation of one-dimensional radiative transfer equation have been used for simulation of spectral radiation intensity. To increase quality of retrieval a spectrum preprocessing and feature extraction method is applied. Simulated spectra were parameterized and expressed as ratios of intensities of multiple rotational lines. Each neural network estimates temperature (NN response) at one point on studied path basing on given spectrum (NN input).