Abstract:Summary Ground roll noise could seriously mask the useful reflection signals and decrease the signal-to-noise ratio (S/N) of seismic data. We propose a fully convolutional framework with dense connections to attenuate ground roll in land seismic data. Our proposed network mainly consists of four blocks, i.e., convolutional, dense, transition down, and transition up blocks. The dense block is consisted of several convolution blocks to extract the waveform features of the seismic data. For the encoder, the transition down block combines the dense block and the 2-D max-pooling layer to compress the model. For the decoder, the transition up block combines the dense block and the 2-D transposed convolution layer to reconstruct the seismic data and restore it to the original size. Besides, the short-long connection in the dense block and the skip connection in the encoder-decoder reuse the features extracted by the previous layer, and add constraints other than the loss function to each convolution block. The well-trained network is tested on two real land seismic datasets containing strong ground roll noise with linear and hyperbolic moveouts, respectively. The testing results show that our proposed method helps to preserve the useful reflection signal and efficiently attenuate the ground roll.

Ground Roll Attenuation Using Convolutional Neural Network with Dense Connected Mechanism