基于粒子群优化的多层介质光热深度剖面重构

Reconstruction of Photo-Thermal Depth Profiles of Multilayer Media Based on Particle Swarm Optimization

基于热波阻抗法建立了多层介质的一维光热模型,从样品表面调制光热信号出发,采用粒子群优化(PSO)算法和总变差(TV)正则化方法对多层介质的热物性予以深度剖面重构。该方法将多层介质离散成一系列厚度相同的虚拟层,并将待重构的热物性深度剖面以粒子表示,再让粒子在解空间进行优化搜索。数值模拟的结果证明了该方法的有效性和实用性,适合用来重构层数未知的多层介质的热导率和热扩散率剖面。对噪声背景下光热信号的重构结果亦证实了算法的稳定性和可靠性。此外,数值结果也表明上述方法可用于热导率和热扩散率深度剖面的同步重构。

A one-dimensional photothermal model of multilayer media is established based on thermal-wave impedance method, and the particle swarm optimization （PSO） method and total variation （TV） regularization are employed to reconstruct the depth profile of thermal properties of the multilayer media by the analysis of the surface photothermal signals. In the proposed method, the multilayer media is discretized into a series of virtual layers with the same thickness, the depth profiles of thermal properties of the multilayer media are represented in the form of particle, and the optimized thermal parameters are obtained by the particles＇ searching in the solution space. The numerical results demonstrate that the algorithm presented in this paper is very effective, and suited to reconstructing thermal conductivity or thermal diffusivity profiles of a multilayer media with unknown number of layers. It is also proved that the algorithm is stable even with noise disturbance. Moreover, the simulation results also show simultaneous profile reconstruction of thermal conductivity and thermal diffusivity is feasible by using the inversion method.