摘要
提出一种基于小波域隐Markov树(HMT)模型的寻北数据去噪新方法。该方法利用HMT模型对信号小波系数的相关性进行建模,通过EM算法对信号小波域的HMT模型参数进行估计,然后进行信号去噪。对实测寻北数据进行去噪表明,该方法比用db4小波阈值法能获得更高的信噪比。
A new signal denoising method based on wavelet-domain hidden Markov tree model （HMT） was proposed, which was applied to data-denofsing for north-finder. The relativity of the wavelet coefficient was modeled by HMT, HMT parameters of signal wavelet-domain were estimated by EM algorithm, and then signal was denoised. The signal denoising results on the measured data show that HMT can obtain much higher PSNR comparing with db4 wavelet thresholding methods.
出处
《弹箭与制导学报》
CSCD
北大核心
2008年第4期245-248,共4页
Journal of Projectiles,Rockets,Missiles and Guidance
基金
国防科技预研基金资助
关键词
小波域模型
EM算法
寻北
信号去噪
wavdet-domain hidden Markov tree model
EM algorithm
north-finder
signal denoising
