基于后向散射模型的多极化SAR影像建筑物高度提取

发布时间：2018-08-23 14:40

【摘要】：以Radarsat-2为例,提出了一种利用多极化SAR影像并结合后向散射模型提取建筑物高度的方法。以北京城区为实验区,首先,分析了SAR影像中建筑物二次散射所对应的亮线连通区域,并统计出该区域对雷达后向散射截面的贡献量;然后,基于平行六面体假设,将建筑物主长度及其与雷达方位向的夹角定量化,并给出计算不同极化散射矢量的方法;最后,利用几何光学-物理光学(geometrical optics-physical optics,GO-PO)模型一阶近似解的后向散射模型估计建筑物高度,并通过比较多个局部训练区提取结果,探讨不同极化信息的提取效果并给出最优极化组合。实验结果表明,通过极化特征组合比仅利用单极化信息提取建筑物高度的精度更高,81.43%建筑物误差小于5 m,均方根误差4.45 m,与ASTER GDEM相关系数为0.909 5,提取结果可靠。
[Abstract]:Taking Radarsat-2 as an example, a method of extracting the height of buildings by using multi-polarization SAR images and backscattering model is presented. Taking the urban area of Beijing as the experimental area, first of all, the bright line connected region corresponding to the secondary scattering of buildings in SAR image is analyzed, and the contribution of the region to the radar backscattering cross section is calculated, and then, based on the parallel hexahedron hypothesis, The main length of the building and the angle between the building and the radar azimuth are quantified, and the method of calculating the different polarization scattering vectors is given. The backscattering model of the first order approximate solution of the geometric optics-physical optics (geometrical optics-physical) model is used to estimate the height of the building. By comparing the extraction results of several local training areas, the extraction effect of different polarization information is discussed and the optimal polarization combination is given. The experimental results show that the accuracy of extracting building height by polarization feature combination is higher than that by using only single polarization information. The error of 81.43% buildings is less than 5 m, the root mean square error is 4.45m, and the correlation coefficient with ASTER GDEM is 0.909. The result of extraction is reliable.
【作者单位】：中国科学院遥感与数字地球研究所;中国科学院大学;
【基金】：国家高分辨率对地观测系统重大专项项目(编号:00-Y30B14-9001-14/16-1、00-Y30B15-9001-14/16-1和00-Y30B14-9001-14/16-2)共同资助
【分类号】：P237

【相似文献】