为解决隧道磁共振超前探测方法只能实现一维深度成像、难以准确反映非层状含水构造空间分布特征的问题，提出一种基于复包络反演的隧道磁共振旋转探测水害隐患的二维成像方法。首先，根据隧道磁共振探测的包络数据，对隧道空间的复包络反演公式和二维复包络反演表达式进行推导； 然后，通过吉洪诺夫正则化、高斯牛顿迭代法和平滑约束条件求取最优化解，得到含水量和弛豫时间2个参数； 最后，以宝兰铁路客运专线麦积山隧道超前探测试验为基础，模拟隧道磁共振旋转探测过程，仿真计算隧道顶板、底板和左右侧壁的含水构造成像结果。结果表明： 1）采用隧道磁共振旋转探测二维成像方法能够在超前探测中得到准确的水害隐患含水量和弛豫时间二维成像结果； 2）磁共振旋转探测二维成像方法预测结果较好，能够准确定位水害隐患位置，预测非层状含水构造的空间分布特征，具有合理性与可操作性。

The conventional advance detection of tunnels using magnetic resonance technology can only obtain onedimensional depth imaging, making it difficult to accurately reflect the spatial distribution characteristics of nonlaminar waterbearing structures and causing safety hazards in tunnel construction. Therefore, using magnetic resonance technology, a twodimensional imaging method based on complex envelope inversion for tunnel rotation detection of water hazards is proposed in this study. First, the complex envelope inversion formula and twodimensional complex envelope inversion expressions are derived for tunnel space based on the detected envelope data of the tunnel using magnetic resonance technology. Then, the water content and relaxation time are obtained by calculating the optimal solution by Gikhonov regularization, Gaussian Newton iteration method, and smoothing constraints. Finally, a case study is conducted on the Maizhishan tunnel of the BaojiLanzhou highspeed railway and the process of magnetic resonance rotation detection is simulated. The obtained imaging results of waterbearing structures in the tunnel roof, floor, left and right sidewalls are numerically calculated. Results of the analysis are the following: (1) The twodimensional imaging results of accurate water content and relaxation time can be obtained in advance detection using magnetic resonance technology. (2) The twodimensional imaging detection results using magnetic resonance are in good agreement with the actual engineering conditions, which can precisely locate the positions of water hazards and predict the spatial distribution characteristics of nonlaminar waterbearing structures, and are reasonable and operational.