| Inhomogeneity of composition,density,temperature of media,and kerr effect,electrostrictive,would cause refractive index continuous and non-uniform distribution,namely,graded index.According to the Fermat principle,ray goes along a curve path within graded index media,and the refraction and total reflection would occur.This makes radiative transfer in graded index media very complex.With the development of scientific research,more and more attention has been paid to the important role of photothermal radiation in many fields such as optical system design,medical imaging diagnosis,atmospheric remote sensing,optical fiber communication,which are urgent to accurately simulate the phot othermal energy transfer characteristics and deeply understand the laws of photothermal information transfer in graded index media.Its core is to study the photothermal information simulation,and optical and thermal parameters identification of graded index media.The essence of photothermal information simulation is radiative transfer solution.The numerical methods based on discrete radiative transfer equation have false scattering and ray effects,and the directional dispersion is not flexible,so it can hardly accurately solve the arbitrary directional radiation intensity information.The accurate calculation of high-resolution radiation information is the premise of the identification of optical and thermal parameters,so it is urgent to develop a new solution technology for solving the radiative intensity along any direction in complex graded index media.The essence of optical and thermal parameters identification is to reconstruct the optical and thermal parameters,such as the graded index,absorption coefficient and scattering coefficient,based on measurement information of photothermal radiative transfer on boundary.However,the parameter identificaiton of graded index media is a typical nonlinear ill posed inverse problem,which involves the cooperative reconstruction of multi parameter field.It needs to solve a series of problems in the reconstruction process,such as ill condition,multi value and so on.Therefore,this paper focuses on the study of the photothermal radiation characteristics and parameters identification of the graded index media,and the main contents include the following five aspects:The numerical solution of photothermal radiation transfer in graded index media is study,and a generalized source term multi flux method based on finite volume is developed.Firstly,the source term distribution in the medium is calculated by the finite volume method,and then according to the direction of radiative intensity to be solved,the ray trajectory is determined by the Runge Kutta ray tracing method.The corresponding radaitive intensity can be obtained by integrating along ray path,so that the directional radiative intensity information of the graded index media can be solved.Based on the principle of integral equation method,an i ntegral equation method based on radiation distribution factor is developed.The radiation distribution factor database is calculated by the reverse Monte Carlo method,and the complex integration problem is solved by the radiation distribution factors,which not only simplifies the integration process of the integral equation,but also retains the accuracy of integral equation.In addition,the database of radiation distribution factors only needs to be calculated once when the physical properties of the medium remain unchanged,so it has high calculation efficiency and can solve the directional radiative intensity information efficiently and flexibly.Based on the theory of swarm intelligent optimization,the invers e model of optical and thermal parameters is established,and the space-dependent and temperature-dependent refractive index and absorption coefficient of semitransparent media are inversed.Two swarm intelligent optimization algorithms(stochastic particle swarm optimization,SPSO and teaching-learning-based optimization,TLBO)are introduced and improved to improve calculation accuracy and efficiency.According to the sensitivity analysis,a corresponding two-step inverse model is proposed to realize simultaneous inversion of refractive index an d absorption coefficient.Swarm intelligent optimization algorithm can’t realize multi-parameter fields cooperative reconstruction and its efficiency is low.Therefore,the reconstruction model of optical and thermal parameter fields is established based on the gradient descent optimization theory.And the refractive index,absorption coefficient and scattering coefficient fields of semitransparent media are reconstructed.Sequential quadratic programming(SQP)algorithm used to solve the smooth nonlinear programming problem is introduced,and the penalty function is adopted to reduce its dependence on the initial solution.Based on the coupled radiation-conduction heat transfer model of graded index media irradiated by laser,absorption coefficient and scattering coefficient fields are reconstructed simultaneously,and absorption coefficient and refractive index fields are reconstructed simultaneously.The hybrid SPSO-SQP algorithm and two-step reconstruction model are developed to realize the simultaneous reconstruction of multi parameter fields.The problem of crosstalk problem exists in the simultaneous reconstruction of multi parameter fields,and the unknown parameters would affect the reconstruciton.Therefore,the model of ray transmission is established based on the theory of ray transmission.The refractive index field of semitransparent media is reconstructed according to the Fermat principle(relationship between ray deflection and graded index).And the absorption coefficient field of semitransparent media is reconstructed according to the Bouguer law(relationship between ray attenuation and absorption coefficient).This reconstruction scheme not only avoids the crosstalk problem of simultaneous multi parameter reconstruction,but also eliminates the influence of unknown parameters,so as to realize the accurate simultaneous reconstruction of the refractive index and absorption coefficient fields. |