| 基于能量运算的磁共振信号尖峰噪声抑制方法 |
| |
| 引用本文： | 万玲,张扬,林君,蒋川东,林婷婷.基于能量运算的磁共振信号尖峰噪声抑制方法[J].地球物理学报,2016,59(6):2290-2301. |
| |
| 作者姓名： | 万玲 张扬 林君 蒋川东 林婷婷 |
| |
| 作者单位： | 吉林大学仪器科学与电气工程学院/地球信息探测仪器教育部重点实验室, 长春 130026 |
| |
| 基金项目： | 国家重大科学仪器设备开发和应用专项(2011YQ030113),国家自然科学基金面上项目(41374075),吉林省科技重点攻关项目(20150519008JH,20140204022GX),中国博士后科学基金面上项目(2014M561296)联合资助. |
| |
| 摘 要： | 磁共振探测信号微弱,使用高灵敏度的核磁共振地下水探测仪,极易受到环境噪声干扰.其中,工频谐波噪声和尖峰噪声,是影响信号质量最严重的两类噪声.国内外研究表明,通过参考线圈的布设,依据探测线圈和参考线圈中噪声相关性,利用自适应参考对消算法,能够有效抑制工频谐波噪声.然而,尖峰噪声的存在严重影响了主通道与参考道的数据相关性,成为了参考对消算法应用的难题与障碍.为解决这一问题,本文提出磁共振信号中尖峰噪声的抑制方法,推导了能量域磁共振信号表达式,通过计算信号能量,可有效检测尖峰噪声并突出不易识别的小幅度尖峰噪声,采用基于中位数的绝对偏差法确定阈值,进而剔除尖峰噪声.为了验证去噪效果,与应用较广的统计叠加法进行对比研究.仿真结果表明,对于干扰幅度较大、持续时间较长的尖峰噪声,能量运算法和统计叠加法均能识别并剔除,且不损失有效的磁共振信号,标准差偏差控制在0.3%以内,可以满足实际应用的要求;对于小于信号幅度1.5倍的尖峰噪声,能量运算法可有效识别并剔除,而统计叠加法无法实现.针对多通道探测系统,使用能量运算法剔除尖峰噪声后,可明显提高主通道和参考道的数据相关性,为后续自适应参考对消算法的应用奠定了基础.实测数据处理结果进一步证明了本文方法的实用性.
|
| 关 键 词： | 磁共振信号 参考对消算法 能量运算 尖峰噪声 噪声相关性 |
| 收稿时间： | 2015-02-02 |
Spikes removal of magnetic resonance sounding data based on energy calculation |
| |
| WAN Ling,ZHANG Yang,LIN Jun,JIANG Chuan-Dong,LIN Ting-Ting.Spikes removal of magnetic resonance sounding data based on energy calculation[J].Chinese Journal of Geophysics,2016,59(6):2290-2301. |
| |
| Authors: | WAN Ling ZHANG Yang LIN Jun JIANG Chuan-Dong LIN Ting-Ting |
| |
| Institution: | College of Instrumentation and Electrical Engineering/Key Laboratory of Geo-Exploration and Instrumentation, Ministry of Education, Jilin University, Changchun 130026, China |
| |
| Abstract: | Magnetic Resonance Sounding (MRS) signal is extremely easy corrupted by the noise, especially by the harmonic noise and spike noise. Harmonic noise cancellation is often based on remote references and the adaptive noise cancellation(ANC)algorithm to increase the signal-to-noise ratio (SNR). However, ANC algorithm cannot play an effective role when spike noise exists. Because the spike noise often lead to a bad correlation between the detection loop and the remote reference loop, it is necessary to remove the spike noise before using the adaptive noise cancellation algorithm to cancel the harmonic noise.#br#In the present paper, we provide a detailed insight into the technique of spike noise cancellation based on transferring the signal from time domain to energy domain and using the median absolute deviation method (MAD). First, we calculate the energy of MRS signal. After doing this, the signal is emphasized which is much more clear than in time domain. Then, a threshold principle based on MAD is provided to cancel all the noise above it.#br#In addition, this paper contains a comparison of the spike noise cancellation effects of Statistical Stacking and Energy cancellation. It is found that the two methods provide identical noise cancellation performance when the spike noise is strong and long enough. Statistical Stacking method is limited when the spike noise get weaker, especially weaker than 1.5 times of the signal. But Energy cancellation we proposed can still remove the spike noise and keep the MRS signal effectively. Moreover, from the noise properties analysis, correlation between the detection loop and the reference loop can be improved after spike noise cancellation. We first apply it on the synthetic data to see the correlation improvement and find that the correlation is improved from 0.1369 to 0.4941 after using the Energy cancellation.Then we obtain the true field data using the multi-channel MRS instrument in Changchun suburb, and apply the Energy cancellation method on the field data. The correlation is improved from 0.1575 to 0.2481. We also find that the harmonic noise is much more evident after spike noise cancellation, which is advantageous to use the adaptive noise cancellation later.#br#We anticipate that better noise cancelling results can be obtained with the Energy cancellation and adaptive noise cancellation methods used together and therefore a wider application of multi-channel MRS instrument can be made in future. |
| |
| Keywords: | Magnetic Resonance Sounding signal Noise cancellation algorithm Energy calculation Spike noise Noise correlation |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《地球物理学报》浏览原始摘要信息 |
| 点击此处可从《地球物理学报》下载免费的PDF全文 |
|
