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

图10 实际地震数据去噪结果 a—真实地震数据;b—DnCNN去噪结果;c—DnCNN去除的噪声;d—b和c的局部相似图;e—Unet去噪结果;f—Unet去除的噪声;g—e和f的局部相似图;h—STCNN去噪结果;i—STCNN去除的噪声;j—h和i的局部相似图

Fig.10 Field seismic signal denoising results a—field seismic data;b—DnCNN denoising result;c—noise data removed by DnCNN;d—similarity map between panels fig.b and fig.c;e—Unet denoising result;f—noise data removed by Unet;g—similarity map between panels fig.e and fig.f;h—STCNN denoising result;i—noise data removed by STCNN;j—similarity map between panels fig.h and fig.i