基于可伸缩型注意力机制的神经网络地震数据去噪方法

张敏, 许一卓, 易继东

A method for seismic data denoising based on the neural network with a retractable attention mechanism

ZHANG Min, XU Yi-Zhuo, YI Ji-Dong

表1 不同噪声水平下的去噪效果性能指标对比

Table 1 Comparison of denoising effect indexes at different noise levels

加入噪声水平 去噪算法 去噪后PSNR/dB 去噪后SSIM σ= 10% PSNR:18.45 DnCNN Unet STCNN 32.91 34.73 35.15 0.9712 0.9837 0.9901 σ= 20% PSNR:12.57 DnCNN Unet STCNN 28.31 29.62 31.12 0.9015 0.9286 0.9474 σ= 50% PSNR:5.36 DnCNN Unet STCNN 13.64 14.12 19.54 0.6845 0.6926 0.7687