针对大地测量型GNSS接收机获取的信噪比(Signal to Noise Ratio,SNR)观测值存在趋势项分离不佳和信号叠加干涉的问题,融合自适应噪声完备集合鲁棒局部均值分解方法对原始SNR信号进行分解,有效分离SNR观测值中的趋势项并提取相应的潮位信号,以布设在美国华盛顿州星期五海港岸边的SC02测站为例,利用实测数据反演潮位变化并与传统的低阶多项式拟合方法进行对比分析.结果表明,融合CERLMDAN方法的反演结果的均方根误差(Root Mean Square Error,RMSE)、相关系数分别为0.56 m和0.97,相比于传统方法的反演精度提升了 33.7％.
Abstract
To address the problems of poor separation of trend terms and signal superposition interference in the Signal to Noise Ratio(SNR)observations obtained by geodetic GNSS receivers,the complete ensemble robust local mean decomposition with adaptive noise method is integrated to decompose the original SNR signals,effectively separating the trend terms in the SNR observations and extracting the corresponding tide level signals.Taking the SC02 station,located in Friday Harbor,Washington,USA,as an example,the tide level variation is inferred from the measured data and compared with the result obtained by the traditional low-order polynomial fitting method.The Root Mean Square Error(RMSE)and correlation coefficient of the inversion results of the fused CERLMDAN method are 0.56 m and 0.97,which improve the inversion accuracy by 33.7％compared with the traditional method.
维普
万方数据