Abstract:Light field (LF) imaging, as an emerging temperature measurement technique, can simultaneously capture the intensity and direction information of radiation in the combustion flow field and has recently attracted extensive attention. However, the dense sampling of the LF introduces a large amount of redundant information and makes the plenoptic data susceptible to noise pollution, which can significantly affect the performance of 3D temperature field reconstruction. To address this problem, a novel approach, i.e., light field compression and noise reduction (LFCNR), is proposed to extract the main features of the projection matrix and separate the noise-related and signal-related subspaces of the plenoptic data, thus improving the reconstruction accuracy and efficiency. The performance of the proposed LFCNR method and the effect of the hyper-parameters on reconstructed quality are investigated thoroughly. It has been observed that the proposed LFCNR-based method for 3D temperature tomography imaging has higher accuracy, more efficiency, and better anti-noise ability compared to the well-established method. LFCNR can improve the reconstructed temperature accuracy by approximately 20%, and the calculation time is shortened to 10%. Priori smoothing is introduced into the optimization objective to constrain the fluctuation of the retrieval temperature, called LFCNR-PS, which can further improve the reconstructed accuracy of axisymmetric and non-axisymmetric fields about 1.79% and 1.23%. Furthermore, experiments of the LFCNR-based method are carried out to reconstruct the 3D temperature field of ethylene diffusion flames. The reconstructed temperature slices demonstrate that the flame topology can be clearly and rapidly identified at different depths. The efficiency and robustness of the proposed method can be applied to the harsh environment, which will be beneficial to the reliable operation of thermal equipment, the study of turbulent chemical reaction mechanisms, fuel utilization, and the optimization design of the engine combustion chamber.

RECONSTRUCTION OF THREE-DIMENSIONAL TEMPERATURE AND CONCENTRATION FIELDS OF A LAMINAR FLAME BY MACHINE LEARNING

Machine learning applied to the retrieval of three-dimensional scalar fields of laminar flames from hyperspectral measurements

Efficient Two-Dimensional Scalar Fields Reconstruction of Laminar Flames from Infrared Hyperspectral Measurements with a Machine Learning Approach

Machine-Learning-Based Reconstruction of Flame Temperature and CO2 Concentration Fields

Studies On Gas Concentration Reconstruction Methods For Simultaneous Temperature And H2o Concentration Tomographic Imaging In A Flame Based On Tunable Diode Laser Absorption Spectroscopy

Machine Learning-Assisted Soot Temperature and Volume Fraction Fields Predictions in the Ethylene Laminar Diffusion Flames.

Machine learning-assisted soot temperature and volume fraction fields predictions in the ethylene laminar diffusion flames

Rapid Tomographic Reconstruction Based on Machine Learning for Time-Resolved Combustion Diagnostics

Experimental Verification of Three-Dimensional Temperature Field Reconstruction Method Based on Lucy-Richardson and Nearest Neighbor Filtering Joint Deconvolution Algorithm for Flame Light Field Imaging

Joint Method for Reconstructing Three-Dimensional Temperature of Flame Using Lucy-Richardson and Nearest Neighbor Filtering Using Light-Field Imaging

Image based temperature field reconstruction for combustion flame

Three-Dimensional Temperature Field Reconstruction from Flame Images Using Algebraic Reconstruction Technique

Numerical and Experimental Reconstruction of Temperature Distribution and Soot Concentration Field for Thin-Slice Flames Using a Charge Coupled Device Camera

Meshed axisymmetric flame simulation and temperature reconstruction using light field camera

Three-dimensional rapid visualization of flame temperature field via compression and noise reduction of light field imaging

Temperature field reconstruction of combustion flame based on high dynamic range images

U-Net Applied to Retrieve Two-Dimensional Temperature and CO2 Concentration Fields of Laminar Diffusion Flames

Three-dimensional Rapid Flame Chemiluminescence Tomography Via Deep Learning.

Reconstruction of 3D Temperature Profile of Radiative Participatory Flame Based on Digital Refocusing Technique of Light Field Camera

Theoretical Fast Non-Intrusive 3-D Temperature Distribution Measurement Within Scattering Medium from Flame Emission Image Analysis