【作者】 周明星；

【导师】 周建江；

【作者基本信息】 南京航空航天大学 ， 通信与信息系统， 2012， 博士

【摘要】 月壤层包含着月球及太阳系物质和能量的大量重要信息。在数亿年的形成和演化过程当中，月壤记录了地月系的起源和演化历史、陨石和微陨石对月球的撞击，以及太阳风和宇宙射线对月球的轰击；反映了月球物质的成分和分布特点。对月壤的系统研究可获得丰富的科学信息。月壤厚度的研究是月壤研究的一项重要的内容，它对以后的月球探测、载人登月、月球基地选址以及开发和利用月球资源等都具有非常重要的意义。在以往的月球探测活动中，月壤厚度一直都是科学家关注的对象。我国第一颗月球探测卫星“嫦娥1号”（CE-1）的科学目标之一就是利用其携带的微波辐射计所获得的月球表面微波辐射亮度温度数据研究月壤厚度的分布，并在此基础上评估月球上3He资源含量。这是国际上首次在月球轨道上用被动微波遥感的方法探测全月球表面的月壤厚度分布。月球表面的微波辐射亮度温度包含着丰富的月球及其外空间的信息，直接反映了月球表层物质的物理化学特性，包括月球表层温度、月表物质的化学组分、密度、介电特性、月壤的厚度、以及地形地貌等。利用从月球轨道上获得的微波亮温数据反演月壤厚度在国际上属于首次，这是一个复杂的过程，目前还有很多问题有待系统的研究和解决。本文主要研究工作和贡献如下：1.研究了月壤中TiO2和FeO的含量、月壤的密度、介电常数分布，以及月球表面粗糙度等多种因素对月球表面亮度温度的影响，为建立月壤中合理的微波辐射传输模型提供了依据。2.研究了月球表面及月壤中温度的分布。在所有影响微波辐射亮度温度的月壤参数中，温度的变化对亮温的变化起主导作用。研究月壤中温度的分布在建立微波辐射传输模型和反演月壤厚度的过程中都至关重要。本文在前人研究的基础上，建立了一维变系数月球表层热传递模型，模拟了月球表面及月壤中温度随时间、月壤深度以及地理位置的变化，并全面分析了各种热物理参数对计算结果的影响，为本文建立月壤的微波传输模型并模拟CE-1号上MRM获得的月表亮温数据提供了依据。3.根据上述研究结果，按照月壤的密度，介电常数及温度随月壤深度的变化，提出了一种非均匀多层的月壤微波辐射传输模型，并分别应用起伏逸散定理和非相关法推导了非均匀多层媒质表面微波辐射亮温与媒质厚度、温度等参数之间的关系，经过比较和化简后得出了统一的亮温计算公式。然后，模拟了CE-1卫星所携带的微波辐射计各频率通道所接收到的月球表面微波辐射亮温随月壤厚度、温度、纬度等变化的规律。与简单分层的模型相比，本文模型能够更加准确地反映亮温与月壤厚度之间的关系，并且模型具有合理性和稳定性。4.研究了CE-1上微波辐射计所获得的2C级亮温数据的处理方法。提出一种基于支持向量机（SVM）和粒子群优化算法的2C级亮温数据回归分析方法，获得了4个频率的微波辐射亮温在月表特定时刻的全球分布，为快速反演月壤厚度分布提供了基础。同时，回归结果也为修正具有明显误差的2C级亮温数据提供了参考。5.根据本文非均匀多层微波辐射传输模型提出了相应的月壤厚度反演方法，并利用3GHz亮温数据反演出了全月球表面月壤厚度的分布。分析表明，本文月壤厚度反演结果具有较高的精度。而且发现月球南北极存在少数极“冷”点，这些地点因为亮温明显小于同纬度平均亮温值而没能反演出月壤厚度。但这些极“冷”点的存在，为研究这些地点是否存在水冰提供了假设前提。6.本文提出利用SVM方法反演月壤厚度。根据纬度、月壤中TiO2+FeO含量、采样时间和月壤厚度一定时各频率通道测得的月球表面亮温之间的关系来训练月壤厚度预测模型。然后利用4个频率的亮温数据和相应的采样信息反演出月壤厚度分布。结果表明，利用SVM方法对月壤厚度的反演结果和利用非均匀多层微波辐射传输模型反演的月壤厚度结果相近，说明基于SVM的月壤厚度反演方法是可行的。更多还原

【Abstract】 Lunar regolith layer contains a large number of important materials and energy information of theMoon and solar system. During the billions of year’s formation and evolution process, lunarregolith not only records the origin and evolution history of the Earth-Moon system, the impacts ofmeteorites and micro-meteorite on the Moon, the bombardments of solar wind and cosmic ray tothe Moon, but reflects the composition and distribution of the lunar material. Plenty of scientificinformation can be obtained by the systematic study of lunar regolith. Study on the thickness of thelunar regolith has very important significance for the future lunar exploration, such as the mannedlanding on the Moon, the lunar base site selection, the exploitation and the further utilization oflunar resources. In the previous lunar exploration activities, the thickness of lunar regolith is allalong concerned by scientists. One of the scientific objectives of China’s first lunar probe "Chang-e1"(CE-1) is to investigate the distribution of the lunar regolith thickness and then to estimate thecontent of3He resource on the Moon by using the microwave brightness temperature data of thelunar surface which were acquired by the carried microwave radiometer. It is the first time usingthe passive microwave remote method on lunar orbit to explore the lunar regolith thickness.The microwave brightness temperature of lunar surface contains plenty information on the Moonand the outer space. It directly reflects the physical and chemical properties of the lunar surfacematerial, including the chemical composition, the density, the dielectric properties, the thicknessof lunar regolith, and the temperature of lunar surface layer. It is the first time that the microwavebrightness temperature data obtained from the lunar orbit are used to inverse the lunar regolith. It isa complex process. There are some remained theoretical problems to be systematically solved. Themain research work and contributions of this paper are shown as follows:1. The influence of various factors on the lunar surface microwave brightness temperature such asthe content of FeO and TiO2in lunar regolith, the density and dielectric permittivity distributionof lunar regolith, and the roughness of the lunar surface are studied, resulting in a basis for thefurther establishment of a reasonable microwave radiative transfer model in lunar regolith.2. The temperature distributions of lunar surface and regolith are studied. In all lunar regolithparameters influencing the microwave brightness temperature, the change of temperature playsa dominant role. Study on temperature distribution in the lunar regolith is very important in theboth processes to establish microwave radiation transfer model and to inverse thickness of lunar regolith. On the basis of previous studies, the establishment of a one-dimensional heattransfer model with variable coefficient for lunar surface layer, the simulation of thetemperature variations of lunar surface and regolith with time, lunar regolith depth, andlocation, and then, comprehensive analysis of the influence of all kinds of thermal physicalparameters on calculated results were carried out in the paper, providing the prerequisite toestablish a reasonable lunar regolith microwave radiative transfer model and to simulate thelunar surface brightness temperature obtained by the microwave radio meter on CE-1satellite.3. Based on the above research results and the variations of density, dielectric permittivity, andtemperature with lunar regolith depth, an inhomogeneous multilayer model is presented.Furthermore, the function relationship between the microwave brightness temperature ofinhomogeneous multi-layered medium surface and the medium thickness, temperature, andother parameters is deduced respectively by applying the fluctuation dissipation theory andnon-coherent method. A uniform formula for calculating the brightness temperature is obtained bycomparing and simplifying process. The changes of lunar surface microwave brightnesstemperatures with time, location and thickness of lunar regolith obtained from four frequencychannels of the microwave radiometer on CE-1satellite are simulated. Compared to the simplylayered model in the literature, the model presented in this paper can more accurately reflectthe relationship between brightness temperature and the thickness of lunar depth. Finally, thestability of this inhomogeneous multilayer model is testified by simulating results based on afine layered model with407layers.4. The data processing method for the2C-level brightness temperature data obtained by themicrowave radio meter on CE-1is studied. A regression and analysis method for the2C-levelbrightness temperature data which is based on Support Vector Machine (SVM) and ParticleSwarm Optimization (PSO) is presented. Then, the distribution of microwave brightnesstemperatures from four frequency channels (3GHz,7.8GHz,19.35GHz,37GHz) on full lunarsurface on specific moments of a lunar day is obtained by using this method. This workprovides the foundation for quick inversion of lunar regolith thickness. Meanwhile, theregression results can also be used as useful references to correct obvious errors in2C levelbrightness temperature data.5. According to the inhomogeneous multi-layer microwave radiative transfer model presented inthis paper, the corresponding inversion method for lunar regolith thickness is put forward, andthen, the full lunar surface regolith thickness distribution is inversed by using the brightness temperature data obtained from3GHz frequency channel. The analysis indicates that thereversion results presented in this paper display the higher accuracy. Additionally the resultsalso indicate that there are a few very "cold sites" lay on the north and south poles of the moon.The lunar regolith thickness of these sites is not inversed because the brightness temperaturesat these sites are significantly smaller than that at the same latitude area. However, thefinding of these very "cold sites" provides the assumptions precondition for the further studieson the existence of water ice at these sites.6. The SVM method inversing of lunar regolith layer thickness is tried firstly in this paper.According to the relationship between the TiO2+FeO content of lunar regolith, lunar latitude,sampling time, and the lunar surface brightness temperature obtained at certain time from fourfrequency channels, the predictive models for lunar regolith thickness has been trained. Then,by using the brightness temperature data obtained from four frequency channels and thecorresponding sample information, the distribution of lunar regolith thickness on the full moonsurface is inversed. The results show that the preliminary results of the inversion of lunarregolith layer thickness based on the SVM method is closed to the inversion result based on theinhomogeneous multi-layer model. It indicts that the method for lunar regolith inversion basedon SVM method is feasible.更多还原