基于RFM模型的叠掩区域定位方法

Positioning Method for Layover Area Based on RFM Model

星载合成孔径雷达(synthetic aperture radar,SAR)是侧视成像,地形起伏较大区域的影像会发生叠掩现象,导致SAR 影像信息丢失,通常采用叠掩补偿解决上述问题。而当前叠掩区域定位方法受数字高程模型(digital elevation model,DEM)精度、相干斑噪声等因素影响,且计算复杂,难以满足快速处理要求。针对上述问题,文章提出一种于基于有理函数模型(rational function model,RFM)的叠掩区域定位方法,首先将DEM 重采样到与SAR 正射影像相同分辨率;随后计算DEM 像元对应的影像坐标;接着统计像方单个像元对应物方单元的个数,大于2 时判定为叠掩区域;最后对生成的掩膜进行编辑,生成最终掩膜。并通过中国“高分三号”卫星SAR 的全极化条带1(full polarized strip 1)QPSI模式获取的两景影像进行试验验证,结果表明,该方法受DEM 影响相对较小,升降轨叠掩试验中补偿率达到99.01%,方法可靠、可行,为SAR 技术在航天遥感应用中提供技术保障。

For SAR images, severe geometric distortions, such as layover, are caused by side-looking viewing geometries and terrain undulations. The layover compensation is usually used to solve the problem. However, the current layover area determination method is affected by many factors such as DEM accuracy and speckle noise, and the calculation is complicated. Therefore it is difficult to meet the fast processing requirements. Aiming at the above problems, this paper proposes a method for layover area based on rational function model (RFM). This method first resamples DEM with the same resolution as SAR orthophotos. Then the image coordinates are calculated corresponding to the DEM pixel. The number of the object cells corresponding to the single pixel of the image side is counted afterwards, and when it is greater than 2, the overlap region is determined. Finally, the generated mask is edited to generate a final mask. The two images acquired by GF-3 satellite SAR QPSI model are experimentally verified. The experimental results show that the proposed method is relatively less affected by DEM, and the compensation rate in ascending and descending experiment reaches 99.01%, which verifies the reliability and feasibility of the proposed method, providing technical support for space remote sensing application.