基于SAR多普勒质心频移的海面流场迭代反演算法

An iterative retrieval algorithm of ocean surface current based on SAR Doppler centroid anomaly

为了克服SAR多普勒质心频移法反演海面流场时风场贡献去除困难的难题,本文提出了基于M4S模型的弦截下山法,利用其迭代计算局部区域的海面流场;然后估算整幅SAR图像中风场对多普勒速度的风贡献因子γ;最后去除风场对多普勒速度的贡献。将该算法用于Radarsat-2数据反演海面径向流速,并利用匹配的实测数据验证反演精度。研究结果表明,本文提出的弦截下山法具有良好的收敛性和较高的收敛速度,而且对本文中使用的两景SAR数据,反演的海面径向流速偏差分别为0.04 m/s和0.15 m/s。

To overcome the difficulty of removing the wind field contribution in the SAR Doppler centroid anomaly method when retrieving the ocean surface current, in this paper, a secant downhill method based on the M4S model is proposed. The contribution of the wind field is removed by iteratively calculating the sea surface current field in a local area and subsequently estimating the wind contribution factor to the Doppler velocity of the whole SAR map. Our method is used to retrieve the radial velocity of sea surface from Radarsat-2 data. Its retrieval accuracy is verified by matching measured data. The results show that the proposed method not only can possess well convergence and high convergent speed, but also the deviation of the radial velocity of the surface is 0.04 m/s and 0.15 m/s respectively for two scene SAR data used in this paper.