The extraction of buildings from remote sensing images is of crucial significance in urban management and planning, but it remains difficult to automatically extract buildings with precise boundaries from remote sensing images. In this paper, we propose the FEPA-Net network model, which integrates the feature extraction and position attention module for the extraction of buildings in remote sensing images. The suggested model is implemented by employing U-Net as a base model. Firstly, the number of convolutional operations in the model was increased to extract more abstract features of the objects on the ground; secondly, within the network, the ordinary convolution is substituted with the dilated convolution. This substitution aims to broaden the receptive field, with the primary intention of enabling the output of each convolution layer to incorporate a broader spectrum of feature information. Additionally, a feature extraction module is added to mitigate the loss of detailed features. Finally, the position attention module is introduced to obtain more context information. The model undergoes validation and analysis using the Massachusetts dataset and the WHU dataset. The experimental results demonstrate that the FEPA-Net model outperforms other comparative methods in quantitative evaluation. Specifically, compared to the U-Net model, the average cross-merge ratio on the two datasets improves by 1.41% and 1.43%, respectively. The comparison of the results shows that the FEPA-Net model effectively improves the accuracy of building extraction, reduces the phenomenon of wrong detection and omission, and can more clearly identify the building outline.

Road extraction from high-resolution remote sesnsing images remains challenging due to complex backgrounds, diverse surface conditions, and the difficulty of maintaining road continuity. Existing U-Net-based methods mainly improve feature fusion strategies, but insufficient feature filtering and limited spatial structure representation often lead to background interference and discontinuous road segments. To address these issues, this paper proposes HFRU-Net, a U-Net-based hierarchical feature regulation framework incorporating an Attention Gate (AG) and a Spatially Enhanced Feed-Forward Network (SEFN). Specifically, AG is integrated into skip connections to adaptively suppress irrelevant background features and enhance road-related representations during multi-scale feature fusion. SEFN is embedded into decoder blocks to strengthen spatial context modeling through multi-scale feature interaction, thereby improving the detection of narrow roads and complex intersections. Extensive experiments on the DeepGlobe and WHU-RuR datasets demonstrate that HFRU-Net achieves mIoU values of 82.31% and 73.28%, respectively, outperforming several existing approaches. Ablation studies demonstrate performance improvements after introducing AG and SEFN, with AG contributing to feature regulation during cross-layer fusion and SEFN enhancing spatial context representation during decoding, while SEFN enhances spatial context representation during decoding, which may facilitate the recovery of narrow and spatially continuous road structures. Furthermore, cross-region evaluation on the GRSet dataset provides preliminary evidence of the transferability of the proposed method. The proposed HFRU-Net provides an effective framework for improving feature representation and spatial structure modeling in remote sensing road extraction.