利用新疆地震台网资料和功率谱密度法，以台站噪声的功率谱密度中值曲线作为评估台站噪声水平的依据，对6个台站的噪声水平进行了评估。对比不同台站的功率谱密度和气枪震源信号的识别情况发现，通过2000次叠加可识别气枪震源信号的台站的噪声水平均低于无法识别的台站，噪声功率谱密度最大差值为40dB，最小差值15dB。最后通过对噪声水平的评估判断，部分台站无法识别气枪震源信号的主要原因是台站噪声水平较高。

We used the data of Xinjiang Digital Seismic Network and PSD(power spectral density)method to do the noise level assessment for six stations. We calculated the median of the power spectral density to evaluate the noise level of different stations. After the comparison of the power spectral density of different stations and the air-gun signal recognition，we found that noise level of the stations with recognizable air-gun source signals is lower than that of the stations without recognizable signals. The largest difference of the power spectral density is 40dB，and the smallest one is 15dB. Finally，we found that the failure to recognize the signal of air-gun source at some stations is due to the noise level.

中国地震局公益性行业科研专项(201508008)、国家自然科学基金(41474051)、新疆维吾尔自治区科研机构创新发展基金(201316)共同资助