煤矿井下地震信号往往呈现出复杂的波场特性且伴随着大量噪音干扰，导致地震信号的初至拾取精度降低，从而影响地震数据的反演与解释。针对复杂干扰环境下采集的低信噪比地震信号，提出了基于变分模态分解(VMD)和遗传算法优化支持向量机(GA-SVM)的地震噪声压制与初至提取方法，以提高煤矿井下复杂噪声条件下的地震信号质量。采用变分模态分解对含噪地震信号进行自适应分解，得到数个的变分模态分量(IMF)；对VMD分解得到的IMF分量进行特征提取，将提取所得的信号特征作为信号有效性判别的依据；利用遗传算法对支持向量机模型进行优化，得到最优的惩罚因子c与核函数参数g；利用优化后的支持向量机模型对IMF分量进行有效性判别并将有效分量重构成高信噪比信号；通过对人工加噪的地震信号应用噪声压制算法，煤矿井下常见的不同类型噪声被有效地压制，验证了算法的可行性；对矿井巷道实采的地震记录进行噪声压制处理，有效地压制了数据中的干扰噪声，极大程度地提高了地震记录的信噪比，使初至拾取得更加准确。结果表明，基于VMD和GA-SVM的地震噪声压制方法可以很好地提取含噪地震记录中的有效信号，提高初至拾取精度，在矿井复杂干扰条件下具有显著的应用潜力，对解决矿井复杂干扰条件下的地震勘探问题有重要意义。

The seismic signals collected underground in coal mines often exhibit complex waveforms accompanied by significant noise interference, leading to a reduction in the accuracy of first arrival time picking of seismic signals and thereby impacting the inversion and interpretation of seismic data. In response to the low signal-to-noise ratio seismic signals collected in complex interference environments, a method for seismic noise suppression and first arrival extraction based on the Variational Mode Decomposition (VMD) and the Genetic Algorithm-optimized Support Vector Machine (GA-SVM) is proposed to enhance the quality of seismic signals under complex noise conditions in coal mines. The approach employs the Variational Mode Decomposition for adaptive decomposition of the noisy seismic signals, yielding several Variational Mode Components (IMF). The feature extraction is applied to the IMFs obtained from VMD decomposition, utilizing the extracted signal features as the basis for discerning signal validity. Genetic Algorithm is utilized to optimize the Support Vector Machine model, obtaining the optimal penalty factor (c) and kernel function parameter (g). The optimized SVM model is then employed for the validity discrimination of the IMF components, reconstructing them into high signal-to-noise ratio signals. By applying the noise suppression algorithm to artificially noised seismic signals, the common types of noise encountered in coal mines are effectively suppressed, validating the feasibility of the algorithm. Noise suppression processing is applied to the seismic records obtained from mine roadways, successfully mitigating interference noise in the data and significantly improving the signal-to-noise ratio of the seismic records, thereby enhancing the accuracy of first arrival time picking. The results indicate that the seismic noise suppression and first arrival extraction method based on the VMD and GA-SVM can effectively separate and extract valid signals from noisy seismic records, improving first arrival time picking accuracy. This approach demonstrates a significant potential for its application in complex interference conditions in mines, which is of significance for the seismic exploration in the mining environments with complex interference conditions.