Calibration Method of Instrument Line Shape for Infrared Radiometer
Sun Yong-Feng,Xu Liang,Shen Xian-Chun,Wang Yu-Hao,Xu Han-Yang,Liu Wen-Qing
DOI: https://doi.org/10.7498/aps.70.20210302
IF: 0.906
2021-01-01
Acta Physica Sinica
Abstract:Interferometric infrared spectral radiometer has high luminous flux and large passing aperture, so the spectral data collected by the instrument are the convolution of the target spectral data and the instrument line shape (ILS). The main factors affecting the ILS include truncation effect, finite field of view effect, off-axis effect, defocus effect and relative position of detector and so on. In this paper the truncation effect, finite field of view effect, off-axis effect and defocus effect on ILS are expounded. These ILS errors all cause the ideal spectrum to drift towards the low wave number and widen. In this paper, the ideal absorption spectrum of water vapor is simulated by the line-by-line integration, through using MATLAB software and combining with the data of water vapor spectrum in HITRAN database, and the quantitative relationship between the error of instrument linear function and spectral distortion is established. According to the simulation results, a factor weight correction algorithm is proposed. It is believed that with this method, the error spectrum caused by ILS has the same optical properties as the ideal spectrum, but the optical path difference is different. Therefore, the coefficient matrix H can be constructed to establish the quantitative relationship between the error spectrum and the ideal spectrum, and the error spectrum can be corrected by using Landweber iterative algorithm. In this paper, the simulation data of water vapor absorption are used to verify the factor weight correction algorithm, and the spectral drift decreases from 0.51 cm–1 to less than 0.01 cm–1. The accuracy of the factor weight correction algorithm is verified by the experimental observation of the standard black body with self-developed interferometric infrared spectral radiometer. The spectral drift of the measured data decreases from 0.226 cm–1 to 0.012 cm–1, and the corrected spectral data are more accurate.