Seismic Stratigraphic Interpretation Based on Deep Active Learning.

Xiaofeng Gu,Wenkai Lu,Yile Ao,Yinshuo Li,Cao Song

DOI: https://doi.org/10.1109/TGRS.2023.3288737

2023-01-01

Abstract:Seismic stratigraphic interpretation plays an important role in geophysics and geosciences. Recently, deep learning has been explored for seismic stratigraphic interpretation. However, deep-learning-based interpretation methods usually require sufficient labeled samples. This is often too hard to be satisfied in field seismic interpretation. In this article, we propose a deep active learning (AL)-based method to address this issue. AL typically exploits prediction uncertainty to reduce labeling effort. We found that uncertainty of prediction is easily obtained in the field of seismic interpretation. Since adjacent seismic images are very similar, they should have similar predictions. When the model performs poorly, the predictions of adjacent images will differ significantly. Thus, the uncertainty can be easily obtained by measuring the similarity of the predictions of adjacent seismic images. Then, data with the highest uncertainty are annotated by geological expert and used for the next round of training. For few-shot AL, initial models obtained by different initial training sets are quite different. We combine deep clustering (DC) and uncertainty sampling to select initial training datasets, with which a good initial model can be obtained. To improve generalization, we introduce a random thin plate spline transformation to simulate changes in terrain. We apply the proposed method to the F3 field seismic data. The results demonstrated that the proposed method can effectively improve the performance of learned seismic interpretation network with very limited labeled samples.