基于同步提取变换的磁共振信号提取方法

Magnetic Resonance Sounding Extraction Method Based on Synchro-extraction Transform

针对磁共振信号(MRS)通常会受到环境中随机噪声影响,导致数据信噪比较低的问题,基于同步提取变换设计了一种MRS提取方法。即在时频分析的基础上,通过同步提取算子得到MRS数据的同步提取变换时频谱,以达到能量集中的目的。基于脊重建原理实现MRS信号的重构。最后,研究了不同噪声水平及不同窗宽对MRS信号提取结果的影响,发现随着噪声水平的增加,信号提取精度逐渐下降;适当增大SET窗函数宽度,可以有效提高MRS信号提取结果精度。

In the practical application,the magnetic resonance sounding(MRS)is usually affected by random noise in the environment,resulting in a low signal-to-noise ratio.To solve the problem,this paper designs a MRS extraction method based on synchro-extraction transform.On the basis of time-frequency analysis,the SET time spectrum of the MRS data is obtained through the synchronization extraction operator to achieve the purpose of energy concentration.Further,based on the principle of ridge reconstruction,the reconstruction of the MRS signal is realized.Finally,we study the influence of different noise levels and different window widths on the MRS signal extraction results.It is concluded that as the noise level increases,the signal extraction accuracy gradually decreases,but appropriately increasing the width of the SET window function can effectively improve the MRS signal extraction accuracy.