Two-dimensional Temperature Field Reconstruction Algorithm of Laser Absorption Spectrum Based on Deep Learning
SHI Yanhui
HAO Xiaojian
HUANG Xiaodong
LI Yunze
WANG Zheng
LIANG Xiaodong
Abstract：This article proposes a temperature field reconstruction algorithm based on a deep learning framework,providing a new method for reconstructing the two-dimensional temperature field of laser absorption spectra.This paper numerically simulates an integrated absorbance array detected from different projection angles,with a specification of N×64(where N=4,8 and 16),and the algorithm constructed in this paper is used to perform temperature field tomography reconstruction on the test set.At the same time,peak signal-to-noise ratio(PSNR),structural similarity(SSIM),and average error are used to evaluate the quality of recon-struction,and the results of two-dimensional temperature field reconstruction for the three modes are studied.The results show that when N=4,the average error of the reconstructed temperature field is 1.32%,the peak signal-to-noise ratio is 37.08,and the struc-tural similarity is 0.961.The reconstruction error decreases with the increase of N value,and the quality of reconstruction improves with the increase of N value,that is,the reconstruction effect is better when N is greater than 4.As the complexity of the tempera-ture field increases,its reconstruction effect relatively decreases,but the reconstruction effect is still good.The research on this al-gorithm is of great significance for the construction of measurement systems and temperature field reconstruction using deep learning methods in actual combustion environments.
Keywords：tunable diode laser absorption spectroscopytwo-dimensional temperature field reconstructionnumerical simu-lationdeep learning
Publication Date：2025-11-20
Online Publishing Date：2026-01-28(First online date of this platform, not the publication date of the document)
Pages：6( 205-210 )