Simultaneous reconstruction of temperature distribution and radiative properties in participating media using a hybrid LSQR–PSO algorithm

Abstract

Chun-Yang(牛春洋) a) , Qi Hong(齐 宏) a) † , Huang Xing(黄 兴) a) , Ruan Li-Ming(阮立明) a) , Wang Wei(王 伟) b) , and Tan He-Ping(谈和平) a)

Geometric calibration of focused light field camera for 3-D flame temperature measurement

“…The intensity detected by the pixel is the intensity of the corresponding ray from the flame. In recent years, various techniques were proposed to reconstruct the 3-D temperature distribution from known radiative intensity of the flames [2] [3]. For instance, Li et al [2] proposed a radiative imaging model based on conventional CCD cameras for flame temperature measurement.…”

“…The 3-D flame temperature field is reconstructed using a Tikhonov regularization method. Niu et al [3] employed a hybrid LSQR-PSO (least-square QR decomposition-particle swarm optimization) algorithm to estimate the 3-D temperature distributions and absorption coefficient simultaneously with known radiative intensity of the flame detected by a hypothetical plane. Several multicameras based tomographic and laser based diagnostics techniques were also reported to reconstruct the 3-D temperature distribution of the flames [4][5][6][7].…”

“…The 3-D temperature reconstruction method used in this study is based on the techniques proposed in [2] [3]. The relationship between the intensity of each ray from the flame and the intensity detected by the sensor can be established by the radiation transfer equations and the model for radiative image formation of the focused light field camera [8].…”

Simultaneous measurement of flame temperature and absorption coefficient through LMBC-NNLS and plenoptic imaging techniques

“…Although the empirical value of the flame radiative properties could be used to reconstruct the flame radiation intensity with the IRT method, the reconstruction accuracy is limited due to the inaccurately assumed radiative properties [7,31]. A better choice for solving this issue is to estimate the 3-D flame temperature distribution and radiative properties simultaneously by using inverse algorithms such as LSQR-PSO [32,33]. However, negative values can be obtained during solving the radiation intensity and properties, and boundary constraints to deal with the negative values have not been investigated [5,32,33].…”

“…A better choice for solving this issue is to estimate the 3-D flame temperature distribution and radiative properties simultaneously by using inverse algorithms such as LSQR-PSO [32,33]. However, negative values can be obtained during solving the radiation intensity and properties, and boundary constraints to deal with the negative values have not been investigated [5,32,33]. Noted that the inverse problem of flame radiation is normally illposed and its solutions are mathematically ambiguous, negative solutions may occur during the required iterations and thus render inaccurate flame temperature reconstruction.…”