Seismic High Amplitude Noise Attenuation Based on the Deep Learning Method
Z. Zhu, D. Cao and B. Wu
Event name: 81st EAGE Conference and Exhibition 2019
Session: Poster: Noise Attenuation A
Publication date: 03 June 2019
Info: Extended abstract, PDF ( 781.38Kb )
Price: € 20
The road traffic noise is a special kind of high amplitude noise in the land seismic acquisition around a road network. We propose a noise detection and attenuation convolutional neural network (NDA-CNN) method, which the conventional detection step with CNN learning method and scale the noisy traces down by further using the standard deviation of the learning result. CNN learning mainly help us to get the complex distribution of high amplitude noises from the noisy data through learning the training data. The distribution includes the mask of noise position and scaler information for next attenuation step, which is more correct than the conventional automatic threshold-determination technique. The real data tests show that NDA-CNN method can get good results for high amplitude noise attenuation, and it is practical and efficient.