Towed transient electromagnetic method (TEM) is an efficient geophysical method for urban underground space measurement. However, compared with the field environment, the power harmonic noise as urban electromagnetic noise is the key reason for the low signal-to-noise ratio (SNR) of the towed TEM data. We propose an efficient preprocessing method based on a multi-step processing strategy, which can fast suppress power harmonic noise and improve SNR. Firstly, a pure power harmonic noise sample is obtained by the least-square fitting. Secondly, the noise component of the measurement data that is most relevant to this sample is quickly searched by an improved genetic algorithm. Finally, the power harmonic noise component is subtracted from the measurement data, and the SNR is significantly improved. The simulation and field experiments show that the new method can fast and effectively suppress power harmonic noise without signal distortion, and it is especially suitable for towed data processing.

拖曳瞬态电磁法（TEM）是一种有效的城市地下空间测量地球物理方法。然而，与现场环境相比，作为城市电磁噪声的功率谐波噪声是拖曳TEM数据信噪比（SNR）低的关键原因。我们提出了一种基于多步处理策略的高效预处理方法，可以快速抑制功率谐波噪声并提高 SNR。首先，通过最小二乘拟合得到纯电力谐波噪声样本。其次，通过改进的遗传算法快速搜索与该样本最相关的测量数据的噪声成分。最后，从测量数据中减去功率谐波噪声成分，显着提高了信噪比。