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8 3种去噪方法在合成地震数据测试中的f-k谱对比 a—干净地震数据的f-k谱;b—含噪地震数据的f-k谱;c—DnCNN去噪结果的f-k谱;d—Unet去噪结果的f-k谱;e—STCNN去噪结果的f-k谱;f—DnCNN去除噪声的f-k谱;g—Unet去除噪声的f-k谱;h—STCNN去除噪声的f-k谱

Fig.8 Comparison of the f-k spectrum of three denoising results in synthetic seismic data testing a—f-k spectra of clean seismic data;b—f-k spectra of noisy seismic data;c—f-k spectra of DnCNN denoising result;d—f-k spectra of Unet denoising result;e—f-k spectra of STCNN denoising result;f—f-k spectra of noise data removed by DnCNN;g—f-k spectra of noise data removed by Unet;h—f-k spectra of f-k spectra of noise data removed by STCNN