The rapid expansion of clean energy systems, particularly wind and solar power, has made large-scale monitoring of related infrastructure using high-resolution multispectral satellite imagery increasingly important yet challenging. Conventional wind turbine detection approaches rely on pixel-level analysis of high-resolution imagery or sliding-window strategies. These methods often suffer from computational redundancy, low efficiency, and high false-positive rates, making them inadequate for supporting large-area surveys and operational monitoring. To address these limitations, this study proposes a Targeted–Refined framework based on AlphaEarth Foundations. First, a wind turbine sample database was constructed for model training and evaluation in Tianjin. In the Targeted Candidate Screening stage, a lightweight multi-layer perceptron classifier exploits the geospatial representations provided by AEF, enabling rapid identification of candidate regions while reducing the search space to approximately 5% of the original image patches. In the Refined Detection stage, high-resolution Jilin-1 satellite imagery is incorporated, and a YOLO model performs precise detection within the candidate regions. Experimental results show that the proposed framework achieves a recall of 96.85%, a precision of 98.41%, and an F1-score of 97.63% on the Tianjin test set. Furthermore, the trained model is directly applied to the plains of the Beijing–Tianjin–Hebei region (excluding Tianjin) for cross-regional evaluation, where it identifies 1,580 wind turbines with an F1-score of 91.33%, demonstrating strong generalization capability. Overall, this framework provides an efficient and scalable solution for detecting large-scale, spatially sparse objects, and offers a practical approach for infrastructure inventory and monitoring of wind farms and similar assets.

The rapid expansion of clean energy systems, particularly wind and solar power, has made large-scale monitoring of related infrastructure using satellite remote sensing increasingly important yet challenging. Conventional wind turbine detection approaches rely on pixel-level analysis of high-resolution imagery or sliding-window strategies. These methods often suffer from computational redundancy, low efficiency, and high false-positive rates, making them inadequate for supporting large-area surveys and operational monitoring. To address these limitations, this study proposes a Targeted–Refined framework based on AlphaEarth Foundations. First, a wind turbine sample database was constructed for model training and evaluation in Tianjin. In the Targeted Candidate Screening stage, a lightweight multi-layer perceptron classifier exploits the geospatial representations provided by AEF, enabling rapid identification of candidate regions while reducing the search space to approximately 5% of the original image patches. In the Refined Detection stage, high-resolution Jilin-1 satellite imagery is incorporated, and a YOLO model performs precise detection within the candidate regions. Experimental results show that the proposed framework achieves a recall of 96.85%, a precision of 98.41%, and an F1-score of 97.63% on the Tianjin test set. Furthermore, the trained model is directly applied to the plains of the Beijing–Tianjin–Hebei region (excluding Tianjin) for cross-regional evaluation, where it identifies 1,580 wind turbines with an F1-score of 91.33%, demonstrating strong generalization capability. Overall, this framework provides an efficient and scalable solution for detecting large-scale, spatially sparse objects, and offers a practical approach for infrastructure inventory and monitoring of wind farms and similar assets.

Land subsidence prediction remains challenging. Conventional grid-based or sequence-only neural networks struggle to represent these spatial dependencies and often lack structured mechanisms for incorporating region-level deformation priors and local physical consistency. This study develops PEGNet, a Peridynamics-inspired and emergent-feature-conditioned spatio-temporal graph neural network for land subsidence prediction. PEGNet uses a per-node GRU for temporal encoding and an EF-conditioned GAT for spatial aggregation. At the node level, historical InSAR deformation and auxiliary hydroclimatic observations are augmented with a PD-derived state and an EF neighborhood context, allowing a per-node GRU to learn temporal evolution and produce a base estimate of future incremental deformation. At the edge level, a fixed spatial graph supports EF-conditioned graph attention, where edge representations combine spatial distance, PD-derived bond-strain information, and the EF same/cross-region gate to generate a spatial residual prediction. At the training and prediction level, the temporal and spatial branches are fused through a learned gate, cumulative deformation is reconstructed from the last observed value, and training combines a data-fitting loss with a PD-inspired local consistency term weighted by distance and EF relations. EF therefore conditions both the network and the coupled regularizer, although it introduces no separate loss term, and the PD component remains a local consistency mechanism rather than a complete Peridynamic solver. PEGNet is evaluated using 60 months of PS-InSAR observations in Tongzhou District, Beijing, four overlapping purged test windows covering nine unique target months and five random seeds. It achieves an incremental-deformation RMSE of 2.153 ± 0.007 mm and a reconstructed cumulative-deformation RMSE of 2.631 ± 0.072 mm. These results outperform the deterministic baselines and remain comparable to the unconstrained GRU-GAT model. Mechanism diagnostics indicate that EF mainly redistributes graph attention, whereas the PD components improve local consistency without producing a substantial global accuracy gain. Overall, the comparable aggregate performance and improved local consistency support the use of PEGNet for conditional subsidence monitoring and scenario analysis.