Fig. 1. Algorithm flow chart

Fig. 2. Rayleigh-BOTDA sensing system

Fig. 3. Three-dimensional Brillouin scattering spectrum

Fig. 4. Front views of noise reduction image with different λ. (a) λ=0; (b) λ=0.5; (c) λ=1

Fig. 5. BFS under different λ. (a) λ=0; (b) λ=1

Fig. 6. Relationship between BFS and λ in normal and heating temperature regions. (a) Normal temperature; (b) heating temperature

Fig. 7. Temperature differences extracted under different λ

Fig. 8. Variation in noise reduction image with number of iterations. (a) n=0, α0=0.9124; (b) n=1, α1=0.2308; (c) n=2, α2=4.2215×10-21; (d) n=3, α3=4.2215×10-21

Fig. 9. Variation in value of ‖un+1-un‖/‖un‖ with number of iterations

Fig. 10. Variation in MSE with number of iterations

Fig. 11. Comparison of TV noise reduction results of 256-time average 3D Brillouin scattering spectrum. (a) Before noise reduction; (b) after noise reduction

Fig. 12. BFS distribution after eliminating under-fitting data