摘要
研究去除多普勒频率估计野值点的方法.在分析野值点统计特性及小波域分布特性的基础上,通过利用野值点和多普勒频率在小波域上系数的不同,首先对各尺度系数进行分段,然后对不同段的系数设定不同阈值进行处理,从而达到去除野值点和恢复多普勒频率的目的.实验结果表明,去除野值点后的结果与原数据的残差中不包含多普勒频率的趋势分量,对后续的脱靶量参数估计没有任何不利影响,去除了随机测量误差与野值点的影响.
A method of removing the outliers of Doppler frequency is presented. Based on the analysis of statistical property of outliers and their distributions in the wavelet domain, different characteristics of wavelet coefficients are found between the outliers and ideal Doppler frequency. On applying threshold method to the wavelet coefficients piecewise, the outliers are successfully removed. Experimental results show the residues between the original and the cleaned data do not contain the component of Doppler frequency. Effects of the stochastic errors and outliers, have been removed and this is helpful to the successive estimation and optimization of miss distance parameters.
出处
《北京理工大学学报》
EI
CAS
CSCD
北大核心
2003年第5期629-632,共4页
Transactions of Beijing Institute of Technology
关键词
小波变换
多普勒频率
野值点
wavelet transform
Doppler frequency
outliers
