🤖 AI Summary
To address inaccurate segmentation of complex geological structures and poor continuity under sparse annotations in seismic horizon interpretation, this paper proposes a Context-Fusion Attention (CFA) U-Net. The model innovatively couples Sobel-based geometric edge priors with spatial attention mechanisms and integrates DBSCAN density clustering to optimize horizon topological connectivity. Additionally, a multi-directional prediction fusion strategy is employed to enhance robustness in fault-proximal zones and folded regions. Evaluated on the Mexilhao field dataset, the method achieves an IoU of 0.881 and a mean absolute error of 2.49 ms. On the North Sea F3 block, using only 5% sparse annotations, it attains 97.6% horizon coverage—significantly outperforming baseline models including U-Net++ and Attention U-Net. The proposed approach establishes a new state-of-the-art in automated seismic horizon interpretation.
📝 Abstract
Interpreting seismic horizons is a critical task for characterizing subsurface structures in hydrocarbon exploration. Recent advances in deep learning, particularly U-Net-based architectures, have significantly improved automated horizon tracking. However, challenges remain in accurately segmenting complex geological features and interpolating horizons from sparse annotations. To address these issues, a hybrid framework is presented that integrates advanced U-Net variants with spatial clustering to enhance horizon continuity and geometric fidelity. The core contribution is the Context Fusion Attention (CFA) U-Net, a novel architecture that fuses spatial and Sobel-derived geometric features within attention gates to improve both precision and surface completeness. The performance of five architectures, the U-Net (Standard and compressed), U-Net++, Attention U-Net, and CFA U-Net, was systematically evaluated across various data sparsity regimes (10-, 20-, and 40-line spacing). This approach outperformed existing baselines, achieving state-of-the-art results on the Mexilhao field (Santos Basin, Brazil) dataset with a validation IoU of 0.881 and MAE of 2.49ms, and excellent surface coverage of 97.6% on the F3 Block of the North Sea dataset under sparse conditions. The framework further refines merged horizon predictions (inline and cross-line) using Density-Based Spatial Clustering of Applications with Noise (DBSCAN) to produce geologically plausible surfaces. The results demonstrate the advantages of hybrid methodologies and attention-based architectures enhanced with geometric context, providing a robust and generalizable solution for seismic interpretation in structurally complex and data-scarce environments.