迭代软阈值法压缩感知重建谱峰较宽的二维固体谱

Compressed Sensing Reconstruction with Iterative Soft Thresholding for Two-Dimensional Solid-State NMR Spectra with Broad Peaks

压缩感知技术可以打破传统奈奎斯特采样定理的限制,利用优化算法对欠采数据进行重建,并获得高质量的结果,因此在核磁共振领域得到了广泛的关注.但是当核磁共振谱的谱峰很宽时,基于共轭梯度方法的压缩感知重建却难以得到令人满意的谱图.因此,该文采用凸优化非线性重建算法,使用基于谱图域软阈值的压缩感知算法重建固体二维宽谱(1H双量子-单量子谱或MQMAS谱),有效地解决了宽峰在重建时变弱的问题.

Conjugate gradient（CG） method has been used previously to reconstruct double quantum-single quantum（DQ-SQ） spectra. However, satisfactory results can be obtained only when the spectra contain only narrow peaks, but not in the case when broad peaks are present. Compressed sensing technology can break the limit of the Nyquist acquisition theorem and reconstruct under-sampled data with high quality. The technology has been applied in the field of nuclear magnetic resonance（NMR）. In this paper, we propose to use compressed sensing with iterative soft thresholding（IST） to reconstruct two-dimensional solid-state NMR spectra with broad peaks, such as the spectra obtained in 1H DQ-SQ and 87 Rb multiple quantum-magic angle spinning（MQ-MAS） experiments. It was found that, with the IST method, compressed sensing technology could be used to reconstruct high-quality spectra containing broad peaks from under-sampled datasets.