混合本征模型的多视SAR影像海冰密度检测

发布时间：2018-06-14 16:26

本文选题：本征模型 + 变异函数　；参考：《中国图象图形学报》2014年12期

【摘要】：目的 SAR影像中像素光谱测度的空间相关性蕴含着海洋表面和海冰更加丰富的空间特性及其变化信息,因此合理建模这种相关性是高分辨率SAR影像海冰精准解译的关键。提出一种利用随机模型及空间统计学测度刻画海冰空间结构的方法。方法本文首先,在空间统计学框架下,SAR影像被表示为多值Gamma模型和泊松线Mosaic模型线性加权构建的混合模型,其中多值Gamma模型用于描述海洋表面雷达信号背向散射变化的连续性,而泊松线Mosaic模型则用于表征不同类型海冰表面雷达信号背向散射变化的区域性。利用上述混合模型的一阶、二阶变异函数,建模蕴含在SAR影像中海冰空间结构的变化。结果对RADARSAT-1影像海冰结构建模并反演其密度。实验区域真实海冰密度分别为20%,80%等,运用本文方法反演所得海冰密度与真实海冰密度误差正负不超过10%。结论本文提出混合本征模型用以刻画SAR强度影像中海冰像素强度变化的空间关系,能够较好地反演Ungava湾海冰密度分布。为利用遥感影像检测空间机构提供一种全新的方法。
[Abstract]:The spatial correlation of pixel spectral measure in SAR images contains more abundant spatial characteristics and change information of ocean surface and sea ice, so it is the key to accurately model the correlation between sea ice and sea ice in high resolution SAR image. Method. Firstly, under the framework of spatial statistics, the SAR image is represented as a mixed model of multi value Gamma model and the linear weighting of the Poisson line Mosaic model, in which the multivalued Gamma model is used to describe the continuity of the backscattering of radar signals on the ocean surface, and the Mosaic model of the Poisson line is used to characterize different types of sea ice. Using the first order, the two order variation function of the mixed model, the modeling contains the changes in the spatial structure of the sea ice in the SAR image. Results in the modeling of the sea ice structure of the RADARSAT-1 image and the inversion of its density. The true sea ice density in the experimental area is 20%, 80%, respectively. The method is retrieved by this method. The error of sea ice density and real ice density error is not more than 10%. conclusion. This paper proposes a mixed eigenmodel to describe the spatial relationship between the intensity changes of sea ice pixels in the SAR intensity image, and can better retrieve the distribution of sea ice density in Ungava Bay, and provide a new method for the use of remote sensing image for spatial mechanism.
【作者单位】：辽宁工程技术大学测绘与地理科学学院遥感科学与应用研究所;
【基金】：国家自然科学基金青年科学基金项目(41301479);国家自然科学基金面上项目(41271435) 中国测绘科学院对地观测技术国家测绘地理信息局重点实验室开放基金项目(201204)
【分类号】：TN957.52

【参考文献】