摘要
为了克服传统X波段雷达测表面流算法存在低流速时误差较大和测量结果不稳定的缺点,提出了一种改进的算法。首先使用色散关系带通滤波器滤除海杂波图像谱中非海浪信号;然后使用调制传递函数对剩余海浪图像谱进行非线性校正,得到真实海浪谱;下一步根据结构风险最小化原则,确定支持向量机中海浪谱分量隶属度;最后使用真实海浪谱和隶属度函数的乘积加权最小二乘法进行初始估流和迭代估流。根据实测岸基X雷达数据实验,结果表明改进算法弥补了传统算法中存在的上述缺点,具有较高的测量精度、稳定性和很好的应用价值。
In order to overcome the traditional current retrieval algorithm's shortcomings based on X-band radar, such as a large error existing at the low current rate and the unsteady result, an improved algorithm is proposed. Firstly, the noise is removed by using a dispersion relationship band filter. Secondly, the remaining non-linear wave image spectra are corrected by means of applying a modern transfer function, and a pure wave spectrum is obtained. The membership of every wave spectral component for support vector machines is determined according to the principle of structural risk minimization in the next step. Finally the product of wave energy and membership function is used as the weight to estimate an initial ocean surface current and carry on estimation iteratively. Using real shore-based X-band radar data for experiments, the results indicate that the improved algorithm has high accuracy and stable results, and it solves the problems in traditional methods. The study provides references for engineering applications.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2012年第2期276-281,共6页
Systems Engineering and Electronics
基金
国家自然科学基金(51009036)
中央高校基本科研业务费专项资金(HEUCFL20111101)资助课题
关键词
海表面流
X波段雷达
色散关系
隶属度
最小二乘法
ocean surface current
X-band radar
dispersion relationship
membership
least square method (LSM)