Flame temperature and emissivity distribution measurement of solid propellant based on clustering analysis of multispectral radiation images

To measure the combustion parameters of a solid propellant, this Letter researches the fitting method for flame temperature and emissivity based on multispectral images and proposes the particle swarm optimization-K-means (PSO-K-means) clustering optimization algorithm of a flame multispectral image. Considering the difference in flame radiation characteristics in different regions, the flame multispectral image is clustered, and spectra in different regions are analyzed and selected in different fitting bands to inverse temperature and emissivity. On this basis, the method is applied to measure solid propellant combustion parameters with different formulations. The measurement shows that the flame temperature is between 1700 and 2100 K, and the emissivity is concentrated in 0.1-0.5. Compared with temperature measurements obtained from tungsten-rhenium thermocouples, the relative deviation of multispectral imaging thermometry is less than 5%. The distribution characteristics of solid propellant combustion parameters with different formulations were analyzed, which provided important data support for evaluating combustion conditions and optimizing solid propellant formulations.