一种神经网络改进小波的地震数据随机噪声去除方法

地震资料的有效信号反射弱，且易受多次波的影响，不可避免地存在随机噪声干扰。提出一种基于神经网络改进小波的地震数据随机噪声去除方法，采用神经网络模型，识别出随机噪声信号，对该信号进行小波包分解，获取多类别随机噪声信号，采用级联BP神经网络模型提取出多类别随机噪声信号，实现地震数据的随机信号压制。实验结果显示，这种改进小波方法对地震数据随机噪声信号的去噪效果较好，在复杂沉积地质结构被探测介质的地震数据随机噪声压制方面具有较强的适用性。

A Method for Random Noise Elimination from Seismic Data Based onthe Neural Network-improved Wavelet Transform

The effective signal of seismic data reflects weakly and is affected by multiple waves, so random noise interference inevitably exists. A method of removing random noise from seismic data based on a neural network-improved wavelet is proposed. The neural network model was used to identify the random noise signal. The signal was decomposed by wavelet packet to obtain multi-class random noise signal. The cascaded back-propagation algorithm (BP) network model was used to extract multi-class random noise signals, the random signal suppression of seismic data was realized. The experimental results showed that the improved wavelet method has a better denoising effect on the random noise signals in seismic data, and has strong applicability in suppressing random noise in the seismic data in complex sedimentary geological structures.