摘要
为了发挥超低场磁共振成像可移动化的优势,传统的被动电磁屏蔽方法应该被主动电磁干扰抑制方法所替代。该文首先分析了电磁干扰传输路径并发现人体耦合为主要传输路径。针对性地设计了用以采集人体耦合电磁干扰的“戒指”型电磁干扰检测线圈,并配合平面及螺线管型线圈共同构成了电磁干扰参考通道。该文提出相关窗口概念:一定采集间隔内的参考电磁干扰信号仍有助于更加精确地估计某一时刻核磁通道电磁干扰信号。选定各通道K空间外围数据为校准数据,由核磁通道校准数据构建校准目标矩阵;由相关窗口内参考通道校准数据构建校准源矩阵。对上述两个矩阵进行小波分解后并通过最小二乘法求得各层转移系数。由参考电磁干扰信号在相关窗口内构建重构源矩阵,并由各层转移系数求得核磁通道电磁干扰拟合值。最后从受污染的核磁信号减去求得的拟合值,实现主动降噪。一维信号以及人体成像实验表明,该主动电磁干扰系统可大幅降低外界电磁干扰对图像质量的影响,使得磁共振扫描仪可在无屏蔽的开放环境中工作。
The ultra-low field(ULF)magnetic resonance imaging(MRI)scanner features much lower cost,much lighter weight,much smaller footprint,and fewer safety concerns than high-filed ones.However,the intrinsic low signal strength of ULF MRI scanners makes them more easily impacted by electromagnetic interference(EMI).The conventional electromagnetic shielding methods passively block the EMI signal but restrict the mobility of ULF MRI scanners.Therefore,the passive electromagnetic shielding should be discarded.The reference channel-based active EMI suppression methods outperform others for better suppression rate and electromagnetic environment adaptation.However,the present deep-learning reference channel-based methods have drawbacks of demanding additional scan time and modification of pulse sequences.This paper proposes a signal-analysis reference channel-based active EMI suppression method with the following features.(1)The EMI detecting coils with appropriate locations avoid sample MR signal.(2)The EMI-detecting coils of reference channels and RF-receiving coils of the MR channel acquire the reference EMI signal and contaminated MR signals synchronously.(3)The reference EMI signals estimate EMI signals in the MR channel(primary EMI signal).(4)The denoised MR signal is obtained by subtracting the estimated primary EMI signal from the contaminated MR signal.The EMI transmission paths are analyzed,and human-body coupling is the main path.The‘ring’shaped EMI pickup coil is designed to sample human body coupled EMI signal.The‘ring’shaped EMI pickup coil,along with one surface coil and one solenoid coil,constitutes three reference channels.It is found that reference EMI signals acquired in a limited time interval also contribute to estimating EMI signals in the MR channel precisely.Then,the concept termed related window is proposed.A data point at a specific time in the k-space of the MR channel is estimated by the data points ink-spaces of all reference channels within the related window.The periphery data of the k-spaces of each channel are selected as calibration data to calculate the transfer matrix for reference EMI signals without the calibration measurement and modification of pulse sequences.The calibration target matrix is constructed from the calibration data of the MR channel,and the calibration source matrix is formed by the calibration data of reference channels in the related window.The two matrixes are decomposed into multiple layers by wavelet analysis.The least-square method calculates the transfer matrix in each layer.The reconstruction source matrix is constructed from k-spaces of reference channels within related windows and decomposed into the same layers.The primary EMI signal is estimated by the reconstruction source matrix and transfer factors in each layer.The denoised MR signal is acquired by subtracting the estimated EMI signal of the MR channel from the contaminated MR signal.The active EMI suppression system is fitted to a home-built portable unshielded 50 mT MRI scanner.The one-dimensional signal analysis and in-vivo scan are carried out in a building located in an industrial park without EMI restrictions.The image results show that the signal-to-noise rate(SNR)is improved by 3.9-fold,and the active EMI suppression system ensures that the ULF MRI scanner works normally without shielding.
作者
杨磊
何为
贺玉成
吴嘉敏
徐征
Yang Lei;He Wei;He Yucheng;Wu Jiamin;Xu Zheng(School of Electrical Engineering Chongqing University,Chongqing 400044 China;Shenzhen Academy of Aerospace Technology,Shenzhen 518057 China;School of Life Northwestern Polytechnical University,Xi’an 710072 China;School of Mechatronics Engineering Harbin Institute of Technology,Harbin 150001 China)
出处
《电工技术学报》
EI
CSCD
北大核心
2024年第15期4708-4717,共10页
Transactions of China Electrotechnical Society
基金
国家自然科学基金(52077023)
重庆科卫联合基金(2023-MSXM016)
深圳市科技创新委员会(CJGJZD20200617102402006)
深圳市科技创新重大专项(KJZD20230923114110019)资助项目。
关键词
超低场
无屏蔽可移动磁共振成像仪
主动电磁干扰抑制
小波分析
相关窗口
Ultra-low field(ULF)
unshielded portable magnetic resonance imaging(MRI)scanner
active electromagnetic interference(EMI)suppression
wavelet analysis
related window
