近地面全偏振光谱成像遥感的高精度定量化反演方法研究

For now the main mode of information processing of polarization spectral imaging remote sensing of the ground objects is using the polarization radiation image processing results of each band, such as polarization degree, polarization angle, etc., to complete qualitative analysis of the ground targets’ polarization features. However, this cannot meet the needs of modern high-precision quantitative remote sensing technology any longer. In this project, research on high-precision quantitative inversion method of full-polarizaition spectral imaging remote sensing closing to the ground is proposed. Firstly, the detection system imaging simulation is completed by modeling of polarized radiation on the entrance pupil, simulation of polarized radiative transfer and modeling of output polarized radiative. Secondly, pBRDF models of typical ground targets are established based on the theory of micro-bins. The models are improved for direct sun light and sky scattered light in the high, medium and low spatial resolution respectively. Then, full polarization vector radiative transfer model is established through the sky light modeling and the establishment of coupling relationship between the ground objects and the sky light. The correction of atmospheric effects are realized in polarization detection. Finally, the accuracy of the inversion of polarization parameters of ground objects is verified according to the full-polarizaition spectral detection experiments closing to the ground. The relative true values of polarization parameters of typical ground objects in different spatial resolution can be obtained and important techniques are provided in order to enhance quantitative level of the polarization imaging remote sensing of the ground objects by this project.

目前对地偏振光谱成像探测信息处理主要方式是将探测系统获取的各波段偏振辐射图像进行处理后得到偏振度和偏振角等信息，完成定性分析以探究地物偏振机理，但这种定性反演已满足不了现代遥感技术的高精度定量化需求。本项目提出近地面全偏振光谱成像遥感的高精度定量化反演方法研究。首先，通过探测系统入瞳偏振辐射建模、偏振辐射传输仿真和输出偏振辐射建模完成探测系统成像仿真。其次，针对典型地物目标建立基于微面元理论的pBRDF模型，在高、中、低三种空间分辨率情况下针对太阳直射光和天空散射光进行模型改进。再次，通过天空光建模和地物-天空光之间的耦合关系来建立全偏振矢量辐射传输模型，实现偏振探测的大气效应校正。最后，完成基于近地面全偏振光谱探测实验的地物偏振特性参数反演精度验证。该项目的研究可精确获取不同空间分辨率下典型地物的偏振特性参数的相对真值，为进一步提高对地偏振成像遥感的定量化水平提供重要的技术手段。

目前对地偏振光谱成像探测信息处理主要方式是将探测系统获取的各波段偏振辐射图像进行处理后得到偏振度和偏振角等信息，完成定性分析以探究地物偏振机理，但这种定性反演已满足不了现代遥感技术的高精度定量化需求。本项目针对近地面全偏振光谱成像遥感的高精度定量化反演方法进行研究。.首先，通过探测系统入瞳偏振辐射建模、偏振辐射传输仿真和输出偏振辐射建模完成探测系统成像仿真。其次，针对典型地物目标建立基于微面元理论的pBRDF模型，可用于反演土壤的含水量、涂层目标的粗糙度，并以涂层为目标，提出了在高、中、低三种空间分辨率情况下针对线性混合像元的空间尺度效应校正方法。实验结果显示，经过校正之后，线性混合像元DOP均值变化不超过0.2%，各尺度下各像元之间DOP的差距有了明显的减少，DOP方差平均减少了13.5%，说明空间异质性的影响降低。.再次，通过天空光建模和地物-天空光之间的耦合关系来建立全偏振矢量辐射传输模型，实现偏振探测的大气效应校正。基于倍加累加法的辐射传输模型和T 矩阵法计算粒子散射特性，建立了适用于多种天气条件下的天空光偏振模型，并提出了基于天空光偏振态分布模型的大气效应校正方法。所建立的天空光偏振模型能够比较准确模拟多种天气下的全天空范围内天空光偏振态分布规律，为不同天气条件下偏振遥感探测提供有力的理论依据。最后，完成基于近地面全偏振光谱探测实验的地物偏振特性参数反演精度验证。实验结果表明，在有云天气下，利用所提出的方法对涂层的粗糙度估算精度为6%，而未经大气效应校正的精度为30.72%，证明了我们所提出的大气效应校正方法对于目标的偏振特性参数反演精度提升的有效性。.该项目的研究可精确获取不同空间分辨率下典型地物的偏振特性参数的相对真值，为进一步提高对地偏振成像遥感的定量化水平提供重要的技术手段。