中文摘要：

本文在考虑半透明介质吸收、发射以及散射衰减作用的情况下采用视在光线法计算某探测方向上介质边界出射辐射强度分布作为逆问题的输入数据,采用LSQR方法在介质辐射特性参数已知情况下重建介质三维温度分布,重建结果显示,无论有无测量误差存在,利用LSQR方法都可以在已知辐射特性参数的条件下很好地重建出介质的三维温度场。在此基础上提出LSQR-SPSO混合算法,并应用于同时重建三维温度场及辐射特性参数（吸收系数、散射系数）。计算结果显示,无论有无测量误差存在,利用LSQR-SPSO混合算法都可以很好地同时重建出介质的吸收系数、散射系数及三维温度场,相比较而言,介质的温度场更容易被重建出来。

英文摘要：

For the semi-transparent media which only consider the absorption,emission and scattering attenuation characteristics,the outgoing radiative intensities at the boundary surface were simulated by the line-of-sight（LOS） method,which were served as input for the inverse analysis.The least-square QR decomposition（LSQR） algorithm was used to estimate the three dimensional temperature distributions of the semi-transparent media with known radiative properties.And the retrieval results show that the temperature distributions could be retrieved accurately even with noisy data.For the participating media with unknown radiative properties,a hybrid least-square QR decomposition（LSQR）- Stochastic Particle Swarm Optimization（LSQR-SPSO） algorithm was developed to estimate the 3-D temperature distributions and radiative properties（absorption coefficients and scattering coefficients） simultaneously.And the retrieval results show that the absorption coefficients,scattering coefficients and 3-D temperature distributions also could be retrieved accurately even with measurement errors.It was also found that the temperature field could be estimated more accurate than the radiative properties.