基于小波域HMT模型的寻北数据去噪方法研究  

Research on Data-denoising Method for North-finder Based on Wavelet-domain HMT

作　　者：邹向阳[1] 陈家斌[2] 徐建华[2] 邹益民[2] 

机构地区：[1]桂林空军学院,广西桂林541003 [2]北京理工大学,北京100081

出　　处：《弹箭与制导学报》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

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.

关 键 词：小波域模型 EM算法 寻北 信号去噪 

分 类 号：TN965.6[电子电信—信号与信息处理] U666.12[电子电信—信息与通信工程]