论文部分内容阅读
合成孔径雷达(Synthetic Aperture Radar,SAR)拥有全天时全天候的工作能力,能够有效地连续对地观测,是土地管理、水体监测、灾害评估等多种应用的稳定数据来源。基于面向对象的思想,提出一种高精度、低虚警率的极化SAR(Polarimetric SAR,PolSAR)水体提取方法。此方法首先对极化SAR图像进行分割,再结合纹理与极化分解特征,对分割区域进行投票,识别水体区域。利用Radarsat-2数据和TerraSAR-X数据开展实验,并将提出方法与基于单一纹理和基于极化分解等水体提取方法进行对比,结果表明该方法在两种数据中均具有最高的总分类精度,其中基于分割技术能够保持完整的水陆边界,纹理与极化特征能够区分浅草、裸地和阴影等与水体相似的地物,结合投票方法能够提高小型水体检测率。
Synthetic Aperture Radar (SAR) has the ability to work around the clock and is capable of observing the earth continuously and effectively. It is a stable data source for many applications such as land management, water body monitoring and disaster assessment. Based on the idea of object-oriented, a Polarimetric SAR (PolSAR) water extraction method with high accuracy and low false alarm rate was proposed. In this method, the polarimetric SAR image is first segmented. Combined with the features of texture and polarization decomposition, the poll area is voted and the watershed area is identified. Experiments with Radarsat-2 data and TerraSAR-X data were carried out, and the proposed method was compared with water extraction methods based on single texture and polarization-based decomposition. The results show that the proposed method has the highest total classification accuracy in both data, Based on the segmentation technique, the complete water-land boundary can be maintained. The texture and polarization features can distinguish similar ground objects such as asak, bare land and shadows, and the voting method can improve the detection rate of small water bodies.