Optical and Acoustic Fusion in Borehole Imaging Logging

Optical Televiewer (OTV) and Acoustic Televiewer (ATV) are cornerstones of establishing borehole imaging logging systems. OTV provides the intuitive analysis of lithology and textures, while ATV has advantages in presenting fractures at deep formation. Combining both imaging methods with integrated interpretation is conducive for geophysicists to characterize structures, and therefore the synthesis based on image fusion is studied in this paper. The self-supervised autoencoder with convolutional neural networks (CNN) and transformer modules is established for feature extraction and reconstruction, where the fusion strategy is implemented in the independent stage. The introduction of the self-attention mechanism makes up for the lack of long-range relationships in CNN. The nested networks capture detailed information in convolutional layers, and the semantic features are conveyed with fine-grained decomposition simultaneously. Besides, the introduction of the self-attention mechanism makes up long-range dependencies for intrinsic information. Quali-tative and quantitative experiments prove that the proposed scheme outperforms competitive methods over textured and structural fidelity.