基于SeaWinds散射计雷达后向散射截面积随风速和相对方位角的变化规律, 从信息图谱几何形态特征出发, 提出了一种新的海面风场反演方法。它不同于传统的依据最小概率误差的遍历搜索算法, 反演时不依赖方差, 仅从信息图谱几何形态特征就可以反演风矢量。实验共用了218623条数据(即218623个风元)。实验结果表明, 与对应的美国NASA反演数据(L2B)对比, 风速平均相对偏差为7.4%, 风向平均偏差为9.5°。

By analyzing the law of Sigma variation with wind speed and relative azimuth, a new approach based on atlas properties is derived in this paper. The approach is different from traditional ones which are ranging methods based on least probability error. Without the need of variance, it can retrieve sea wind just by atlas information. 218623 data were in the experiment, and it turned out that the wind vector result was close to L2B data. Mean relative error of wind speed is 7.4%, and mean error of wind direction is 9.5°.