Fig. 1. Flow chart of broadband temperature measurement algorithm and process of spectral fitting optimization. (a) Flow chart of broadband temperature measurement algorithm; (b) process of spectral fitting optimization

Fig. 2. Normalized line strength and lower state energy distribution of 79 absorption lines at temperature range of 300-2000 K.(a) Normalized line strength; (b) lower state energy distribution

Fig. 3. Fitting result of test spectrum and raw spectrum with ±0.005 amplitude Gaussian noise and ±0.005 amplitude Gaussian noise.（a）Fitting result of test spectrum and raw spectrum with ±0.005 amplitude Gaussian noise；（b）Gaussian noise with amplitude of ±0.005

Fig. 4. One standard deviations and error bars of 700 K, 1200 K, 1700 K calculated at Gaussian noise with amplitude from ±0.005 to ±0.1. (a) One standard deviation; (b) error bar under temperature of 700 K; (c) error bar under temperature of 1200 K; (d) error bar under temperature of 1700 K

Fig. 5. Fitting result and residual of raw spectrum and test spectrum under temperature of 800 K. (a) Fitting result; (b) residual

Fig. 6. One standard deviation and maximum absolute deviation of temperature calculation results under different error perturbation of S₀, γ_self, γ_air, δ_j and n_air. (a) One standard deviation under temperature of 1000 K; (b) one standard deviation under temperature of 1300 K; (c) maximum absolute deviation under temperature of 1000 K; (d) maximum absolute deviation under temperature of 1300 K

Fig. 7. One standard deviation of temperature calculation results under combination perturbation of spectral noise and spectral parameter error. (a) 1700 K; (b) 800 K