GNGTS 2023 - Atti del 41° Convegno Nazionale

Session 3.3 ______ ___ GNGTS 2023 Figure 3: Field data result. (a) Original fully sampled data. (b) 50% regularly undersampled data. (c) Interpolated result using the presented method. (d) Error obtained using the presented method. The field data is from Gulf of Suez. It has 128 trace and 512 time-samples per trace. The temporal sampling interval is 4ms, and the spatial interval is 12.5m. Figures 3a and 3b, respectively, display the original fully sampled data and the regularly undersampled data. The interpolated results utilizing our suggested methodology with SNR=9.12dB are shown in Figure 3c. We can notice that Figure 3c's events are more continuous than 3b's. The corresponding errors of the presented method are shown in Figure 3d, which contains little useful signal, which proves the great performance of our presented method. CONCLUSIONS We presented an end-to-end self-supervised deep learning method based on the equivariant imaging system for regularly undersampled seismic data interpolation. With the help of the fact that seismic data is equivariant with respect to shift and undersampling, a training strategy that enforces both the measurement consistency and the shift transform equivalence is utilized such that the deep CNN can capture more prior information of seismic data. The performance measured on field example demonstrate that our presented method has great interpolation performance. ACKNOWLEDGMENTS Graphics processing unit (GPU) computation has been made available, thanks to the GPU Grant Program, by NVIDIA Corporation, to which the authors express deep gratitude. REFERENCES Chen, D., J. Tachella, and M. E. Davies, 2021, Equivariant imaging: Learning beyond the range space. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 4379-4388. Chen, Y., M. Bai, Z. Guan, Q. Zhang, M. Zhang, and H. Wang, 2019a, Five dimensional seismic data reconstruction using the optimally damped rank-reduction method: Geophysical Journal International, 218, 224–246. Fang, W., L. Fu, M. Zhang, and Z. Li, 2021, Seismic data interpolation based on u-net with texture loss: Geophysics, 86, no. 1, V41–V54. Fomel, S., 2003, Seismic reflection data interpolation with differential offset and shot continuation: Geophysics, 68, 733–744. Hennenfent, G., and F. J. Herrmann, 2008, Simply denoise: Wavefield reconstruction via jittered undersampling: Geophysics, 73, no. 3, V19–V28.