【摘要】 为了尽可能地剔除噪声、保留有用信号，提出了一种自动编辑地震道的两步法，该方法先在局部时窗内根据采样点的振幅与平均振幅的比值对地震道中的尖脉冲和猝发噪声作精细切除，然后采用一个３层ＢＰ人工神经网络对地震道进行整过编辑，以剔除其他噪声。实际资料试处理效果分析表明：该方法不漏切噪声，也不会将有用信号盲目丢失。更多还原

【Abstract】 The objective of editing seismic data is to remove seismic noises and makes seismic signals reserved. An effective method for automatic editing seismic data through two steps is proposed for doing this. This method firstly removes pulse and burst noise within smaller time windows and then other noises are removed by using an artificial neural network. Results processed for real seismic data reveal that this method can reserve seismic signal and remove seismic noise from seismic data as much as possible.更多还原