Research Of Imaging Methods For Laminar Flame Based On Single-Pixel Camera

Combustion is the dominant form of energy conversion for a span of power systems such as engines and power plant boilers,so it is of strategic significance for human beings.Increasing efficiency of combustion by conducting combustion diagnostics and studying combustion mechanisms is essential to the whole society.However,combustion is an extremely complicated process which produces a huge number of intermediate products that usually feature non-uniform spatial distributions,bringing great challenges to combustion diagnostic methods.Imaging methods such as planar laser-induced fluorescence or chemiluminescene imaging can obtain abundant information of flame by measuring free radicals emitting in various bands and help researchers to understand combustion mechanisms better.Imaging methods usually acquire array sensors with high sensitivity,and broad bandwidth.However,array sensors may be quite expensive in some bands,or even no commercial product is available in this band.Besides,image quality of array sensors deteriorates drastically when the wave length of target approaches the size of sensor pixels.The actual intensity of signals might be weak in some bands,so the sensitivity and signal-to-noise ratio of array sensor may not meet the requirements.Single-pixel detector shows the advantages of low-cost,broad available bandwidth and high sensitivity and have great potential of overcoming problems of array sensors mentioned above.To investigate the possibility of implementing single-pixel imaging to combustion diagnostics,this paper report the attempt of constructing a single-pixel camera based on compressive sensing theory.Simulations have been conducted to determine the optimized parameters such as sampling ratio,measurements matrix and reconstruction algorithm.A validation as well as demonstrative experiment was also conducted and the results show high similarity to those of CMOS camera,indicating the feasibility of this single-pixel camera for combustion diagnostics.For the problem of the relative low temporal resolution,this paper made attempt to optimize this by using computed tomography based on sparse-view data.The simulation results show that this method can reduce sampling ratio needed and increase the temporal resolution.Besides,results of validation experiment shows better quality than traditional methods,proving the method proposed in this paper can reach higher temporal resolution while keeping satisfactory image quality,so the singlepixel camera based on computed tomography of sparse-view data can be utilized in combustion diagnostics.