基于经验模态分解和小波分析的靶船外弹道测量数据去噪方法

Target Ship Exterior Ballistic Data Denoising Method Based on Empinical Mode Decomposition and Wavelet Transformation

针对舰船外弹道测量数据误差性质复杂,表现为非平稳信号,难以用模型准确进行描述的问题,提出基于经验模态分解和小波去噪的靶船外弹道测量数据处理方法。首先对测量数据进行经验模态分解,得到测量数据不同频率的本征模函数,接着采用小波间值去噪原理对含有噪声的高频本征模函数去噪,最后将经过去噪的本征模函数与剩下的没有经过去噪的本征模函数和趋势项相加,重构去噪后的靶船测量数据。经实验数据分析,本方法在很大程度上克服了小波阈值降噪的缺陷,保留了高频分量中包含的有用信息,是分析非平稳、多频段的靶船外弹道测量的有效方法。

The complexity of errors of target ship exterior ballistic is characterized by its non-stationary signals, which makes it fairly difficult to be described accurately by mathematical models. In order to solve this problem, a method based on empinical mode decomposition（EMD） and wavelet transformation is proposed. Firstly, EMD is applied to the data to get the intrinsic mode functions（IMFs） of different frequencies, then the denoised IMFs is added to the rest together with the trend to reconstruct the measuring data of the target ship. The experimental results show that this method can enormously improve the performance of traditional wavelet threshold method and remain the useful information underlying the high frequency data, thus the effectiveness of this method in dealing with the non-stationary and multi-frequency data of target ship exterior ballistic is effective.