基于分数阶矩和NM单纯形算法的海杂波参数估计  

Parameters Estimation of Sea Clutter based on Fractional Order Moments and NM Algorithm

作　　者：杨永生[1] 张宗杰[1] 

机构地区：[1]苏州科技学院电子信息工程学院,江苏苏州215011

出　　处：《遥感技术与应用》2011年第6期771-774,共4页Remote Sensing Technology and Application

基　　金：江苏省高校自然科学研究项目(11KJB420003)资助

摘　　要：针对海杂波K分布形状参数和尺度参数的估计问题,提出了基于分数阶矩和NM单纯形算法的参数估计,首先采用分数阶矩估计法作为形状参数和尺度参数的初始值,然后利用NM单纯形算法进行数值搜索获得K分布形状参数和尺度参数的估计值。与实测的X波段海杂波概率密度函数对比表明,K分布优于Gamma分布和Weibull分布,其估计值与实测值最相吻合,且具有最小的均方根误差。对于一维的海杂波来讲,该方法是一种计算效率适中的算法。Aiming at the question of shape and scale parameters estimations of K-distribution,a method combing fractional order moments with NM simplex algorithms is presented.Initial values of shape and scale parameters are calculated via fractional order moments,and then the estimated values are obtainded by numerical searching algorithms of NM simplex.We have analyzed measurement data of X-band sea clutter.The results are shown that the K-distribution is superior to Gamma and Weibull distribution.The probability density function of K-distribution is consistent with the measurement data,and it has the lowest RMSE.As far as one dimension sea clutter is concerned,the method has moderate computation efficiency.

关 键 词：Nelder-Mead(NM)单纯形算法 海杂波 K分布 参数估计 

分 类 号：TN959.72[电子电信—信号与信息处理]