Adaptive spectral band identification method in noncontact temperature measurement based on dynamic programming

Abstract

An adaptive method based on dynamic programming is proposed to identify the spectral band for non-continuous measurement of surface temperature of heatshield materials when Fourier transform infrared (FTIR) spectrometer is used to collect the radiation spectrum in the dynamic heating environment of a high-frequency plasma wind tunnel. First, the radiation spectrum is converted to a time series. Then, high-frequency parts of the measurement spectral signal are obtained by multi-resolution analysis of one-dimensional discrete wavelet and then the suitable spectral band required by a non-contact temperature measurement is adaptively identified based on dynamic programming. Eventually, surface temperature and its corresponding emissivity can be determined. Results of the experiment conducted on a benchmark material (graphite) in the dynamic heating environment of high frequency plasma wind tunnel show the proposed method to be practical.