散射光谱的目标材质及比例反演研究

研究了基于散射光谱的点目标在远距离条件下表面材质及比例反演方法，旨在为空间碎片探测及预报提供数据参考。首先根据散射光谱的远距离目标探测物理模型，建立了基于散射光谱的目标参数反演物理模型，给出了基于最小二范数理论方法的目标表面材质及比例信息的反演算法，结合光源照射特性、目标材料表面光学反射特性、入射、反射及探测角等信息，利用双向反射分布函数(BRDF) 的多级融合模型，表征复杂材料表面的光学反射特性, 将BRDF中对应的面积量作为待反演参数，给出了目标表面材质及比例信息的反演算法；其次进行了实验验证，搭建了室内散射光谱探测及采集系统，进行了单一材质及多种材质不同比例的目标散射光谱探测及数据采集，经过对散射光谱数据进行预处理，截取有效波长范围为400～800 nm；结合理论分析及反演算法，对于四种材质组合的样品，进行了相同和不同比例组合的材质及比例反演，等比例反演结果最小误差为0.8%。最大误差为13.6%，平均误差为4.9%；不等比例反演结果最小误差为6%，最大误为12%，平均误差为9.25%；综合以上测试结果可以得出，反演平均误差最大为9.25%，考虑到入射光源稳定性有2.89%的误差，反演结果最大平均误差将小于6.36%。得到了平均反演误差小于10%的良好效果，验证了该方法的准确性；最后以某失效卫星为例，对其在轨的散射光谱信息，进行了材质及比例反演计算，给出了其主要材质为太阳能帆板、保温膜及碳纤维板，与真实情况相符较好。该研究为点目标远距离材质比例反演识别提供了新的技术途径。

Inversion of Object Materials and Their Proportions Based on Scattering Spectra

The paper investigated the inverse method for the surface materials and the proportions of the space object from long distances based on the scattering spectra. The research results shall provide the data references for space debris detection and forecasting. Firstly, based on the long-distance object detection physical model of scattering spectra, we first constructed the object parameter inverse physical model based on scattering spectra and provided the inverse algorithm for object surface material and its proportion based on the least norm theory method. By combining with the lighting characteristics, object material surface optical reflection characteristics, incidence, reflection and detector angle data, etc., using the multimodal fusion model of the bidirectional reflectance distribution function (BRDF) and characterizing the optical reflection characteristics of the complex material surface. We took the corresponding area in the BRDF as the parameter to be inverted and obtained the inversion algorithm of the object surface materials and its proportion information. Secondly, we performed the experimental validation by building an indoor scattering spectrum detection and acquisition system to perform the scattering spectrum detection and data acquisition of single material and multiple materials with different scales. We intercepted the effective wavelength range of 400~800 nm by preprocessing the scattering spectrum data. Combined with theoretical analysis and inversion algorithm, we took the samples with 4 kinds of materials and performed the material and its proportion inversion with equal proportional and non-equal proportion combination for the samples. The minimum, maximum and average errors of the equal proportion inversion are 0.8%, 13.6% and 4.9%. Moreover, the minimum, maximum and average errors of the non-equal proportion inversion are 6%, 12% and 9.25%. Therefore, according to the above testing results, the maximum average error for inversion is 9.25%. Once considering the error influence of 2.89% from the incidence light source, the maximum average error for inversion will be lower than 6.36%. So we suggested the average inversion error will be less than 10%. Thereby the accuracy of the inverse method is verified. Finally, take one failed satellite as an example. The proposed method was used to invert the materials and proportions of its on-orbit scattering spectra. Its surface materials consist of solar arrays, insulation film and carbon fiber plates, which matches well with the real situation. In summary, this paper has provided a new technical approach for the inversion identifying of space object materials and their proportions from long distances.