摘要
核磁共振的信噪比在T2谱反演中起着重要作用,但是常规的估计方法依赖于奇异值的截断值选取以及控制方程组的系数矩阵,因而给噪声方差的估计带来一定的任意性,并且不能自适应于信号的变化。提出了一种小波估计方法,该方法估计的噪声方差仅仅与信号有关,与控制方程组无关,并且能够自适应于信号的变化,由其给出的噪声信号具有高斯分布特性。
The SNR plays an important role in the inversion of NMR＇s T2 spectrum. The conventional methods for SNR estimation are very much dependent on the cutoff of the singulars and the coefficient matrix of the inversion system. Thus inevitably introduced uncertainty in the SNR estimation. We present a new method for the SNR estimation by wavelet estimation. The results from this method are only dependent on the signal. Furthermore, this method can adapt to the variations of signals. Our results also demonstrate that the nature of the NMR noise signal is of Gaussian distribution.
出处
《核电子学与探测技术》
CAS
CSCD
北大核心
2005年第5期471-474,共4页
Nuclear Electronics & Detection Technology
