一种基于噪声统计特性的改进EMD降噪方法

GPS高程时间序列中通常都含有噪声，容易对GPS信号解算精度造成影响。针对这一问题，本文基于噪声统计特性，提出了一种改进的EMD降噪方法。首先将信号进行EMD分解，得到低频信号与高频噪声两个部分；然后将高频噪声部分随机打乱两次，并与原始高频噪声累加，求取平均值；最后与低频信号累加，构成一个新的信号再次进行EMD分解，提取出有用信号。最终利用模拟数据和WUHN站实测GPS高程时间序列数据对该方法进行验证。试验结果表明，当信噪比较高时，本文方法得到的降噪效果更佳。

An improved EMD noise reduction method based on noise statistical characteristics

The GPS elevation time series usually contains noise, which easily affects the accuracy of GPS signal solution. In view of this problem, this paper proposes an improved EMD noise reduction method based on the statistical characteristics of noise. Firstly this method decomposes the signal by EMD to obtain two parts of low-frequency signal and high-frequency noise. Then randomly shuffles the high-frequency noise part twice and accumulates with the original high-frequency noise to obtain the average value. Finally, it compares with the low-frequency signal add up to form a new signal and perform EMD decomposition again to extract the useful signal. Verification is performed using simulated data and measured GPS elevation time series data from WUHN stations. The experimental results show that when the signal-to-noise ratio is relatively high, the noise reduction effect obtained by this method is better.