基于北斗反射信号的支持向量回归机海面风速反演研究

随着全球导航卫星系统的发展,一种基于导航卫星反射信号(GNSS-R)的海洋遥感技术成为近年来研究的热点。为研究GNSS-R测风方法,分析了海面风速与反射信号时延功率之间的关系,同时结合机器学习方法中的支持向量回归机方法,将反射信号时延功率作为训练标签,最终完成海面风速的反演。将"北斗反射信号海风海浪探测系统——山东示范站"项目期间海风海浪系统的探测结果与同期风速仪探测结果进行对比分析,结果表明,两者具有很好的一致性,海面风速的反演精度为2.4 m/s,证明了基于北斗反射信号的支持向量回归机海面测风方法是可行的。

Study on Sea Surface Wind Speed Retrieval Based on Support Vector Regression Machine Using Beidou Reflected Signals

With the development of Global Navigation Satellite System,a marine remote techniquebased on Global Navigation Satellite System reflected signal (GNSS-R) has become a hotspot in remote sensing fields. This paper focuses on wind speed retrieval based on GNSS-R. Firstly, the relationship between sea surface wind speed and delay power of reflected signals is analyzed.Then this paper employs support vector regression machine (SVRM) to retrieve the wind speed, with the delay power of reflected signals being the training label.In the end, this paper emphasizes comparing the Beidou Satellite (BDS) reflected signals detected by ocean wind and wave system with the wind speed data from anemographare during the experiment in BDS reflected signal ocean remote sensing station of Shandong. It shows a good consistency, and the retrieval accuracy of sea surface wind speed is 2.4 m/s. It proves the feasibility on detecting wind speed through SVRM method using BDS reflected signal.