高分3号星载合成孔径雷达极地海冰自动检测方法研究

随着全球变暖等一系列气候变化的发生，极地海冰成为人们日益关注的焦点。由于不受光线和云雨影响，合成孔径雷达（SAR）可以进行全天时全天候的观测。高分3号是我国高分系列卫星中的一颗星载合成孔径雷达成像卫星，具有多种成像模式，可以在全球获取SAR数据。全天时全天候的工作特性和高空间分辨率的优势，使得高分3号星载SAR在极地海冰遥感监测中发挥重要的作用。本文基于高分3号水平-垂直（Horizontal-Vertical，HV）极化数据，提出了一种基于支持向量机的无需人工干预的海冰检测方法，实现海水和海冰的自动分离。利用该方法得到的海冰和海水分离结果同辅以人工解译的半监督分类结果相比较为吻合，为高分3号服务于极区海冰监测奠定了良好的基础。

The method study on automatic sea ice detection with GaoFen-3 synthetic aperture radar data in polar regions

With climate change occurrence, such as global warming, polar sea ice has been drawn increasing attentions. Synthetic aperture radar (SAR) can monitor earth independent on sunlight and cloud. GaoFen-3 is a C-band SAR of the GaoFen series satellites, which has multiple imaging modes and can obtain data globally. SAR plays an important role in monitoring polar sea ice due to the advantage of all-weather operation and high spatial resolution. Bases on the GaoFen-3 horizontal-vertical data, we proposed an automatic method to discriminate sea ice and sea water, using the support vector machine classification method. The detected sea ice shows good agreement with visual inspection. The successful development of this algorithm supports the Gaofen-3 SAR for operational service of monitoring polar sea ice.