斑点抑制与多分辨率拓扑分析相结合的SAR图像河流水体提取

针对由于斑点噪声和建筑遮挡造成的高分SAR图像河流水体提取的不连续问题，本文提出一种结合自适应斑点抑制和多分辨率拓扑分析的河流水体提取方法。首先，在斑点抑制各向异性扩散中引入Beta度量以实现SAR图像的自适应滤波；其次，利用局部阈值分割和连通域标识对去噪图像中的河流水体段进行快速提取；然后，以河流水体段的最小外接凸多边形为对象构建高斯金字塔，并由此进行多分辨率拓扑分析来获得河流间断点信息；最后，由区域生长策略逐层地对间断点进行判别，实现相邻河流水体段的自动连接。采用TerraSAR-X和GF-3高分SAR图像验证提出的方法。结果表明，该方法可有效地提取完整河流水体。对多幅SAR图像提取结果的定性和定量评价验证了提出方法的有效性与稳健性。

River waterbody extraction from SAR images based on speckle reduction and multi-resolution topological analysis

Aiming at the discontinuity of river waterbody extraction from high-resolution SAR images caused by speckle noise and building occlusion,a river waterbody extraction method combining adaptive speckle reduction and multi-resolution topological analysis is proposed.Firstly,the Beta metric is introduced into the speckle reduction anisotropic diffusion to achieve adaptive speckle filtering of SAR images.Secondly,the local threshold segmentation and connected region identification are used to quickly extract river waterbody segments from the denoised image.Then,a Gaussian pyramid is constructed with the smallest circumscribed convex polygon of the river waterbody segments as the object,and multi-resolution topological analysis is performed to obtain the information of river discontinuities.Finally,the regional growth strategy is used to distinguish the discontinuities layer by layer to realize the automatic connection of adjacent river waterbody segments.The proposed method is verified by TerraSAR-X and GF-3 high-score SAR images.The results show that the proposed method can effectively extract complete river waterbody.The effectiveness and robustness of the proposed method are verified by the qualitative and quantitative evaluation of multiple SAR images.