Road extraction from remote sensing imagery is a crucial task in fields such as intelligent transportation. However, challenges remain in restoring road details in complex scenes, multi-scale modeling, and ensuring computational efficiency. To address these issues, this study proposes AFDANet, whose core components include: (1) The Spatial-Enhanced Adaptive Feature Fusion (SEAF) module, which dynamically integrates shallow detail features with deep semantic features. This module incorporates a novel Parallel Channel-Spatial Attention (PCSA) mechanism. PCSA effectively decouples channel and spatial information, independently modeling semantic and spatial dimensions, thereby enhancing the model’s ability to distinguish roads from complex backgrounds. (2) The Strip Grouped Aggregation Decoder (SGADecoder), which combines multi-scale multi-directional strip convolutions with residual connections, effectively restores road continuity and edge details. Experimental results on the DeepGlobe and Massachusetts datasets demonstrate that AFDANet outperforms existing advanced methods across multiple key metrics while requiring fewer parameters and lower computational costs. This study provides a high-accuracy and computationally efficient solution for road extraction from remote sensing imagery. The source code is publicly available at https://github.com/ZehuaChenLab/AFDANet.

High-resolution satellite imagery for road extraction plays a crucial role in urban planning and geographic information updates. However, the discontinuity and breakability of extracted road images pose challenges to extraction methods. Additionally, the complexity of the remote sensing imagery background can lead to interference from similar objects in the surrounding environment. To alleviate these problems, we propose a Context-Aware and Road-Enhancement road extraction network (CARENet). To enhance the continuity and integrity of the extracted roads, a bidirectional strip feature extraction module (BSFEM) is designed in skip connections. This module is a novel strip feature extraction method that can preserve edge information of the roads at each scale and pass it to the decoder, providing rich and accurate road detail features. Subsequently, a dilated conv-based Selective Scan module (DBSSM) is designed to achieve linear attention while minimizing the negative effects of complex backgrounds. The DBSSM consists of multiple context-aware blocks using 2-D Selective Scan (SS2D) to capture contextual relationships. Experiments conducted on two public road datasets demonstrate that CARENet outperforms several recent methods in various evaluation metrics, including Intersection over Union (IoU) and${F}1$-score. Our source code is available athttps://github.com/ZehuaChenLab/CARENet.

Road extraction from remote sensing imagery plays a pivotal role in a wide range of geospatial and urban applications. Nevertheless, this task remains inherently challenging due to the intricate morphological variations of roads and frequent occlusions or interference caused by complex background environments. To address these challenges, we propose a road extraction network based on Gated Global-Local Linear Attention (G2L2Net). Firstly, we introduce a linear deformable convolution and design a Linear Input-Dependent Deformable Convolution (LID2Conv) which adaptively modulates convolution offsets and weights in a content-aware manner. In addition, we design a Top-K based sparse gated weight (TGW). We use this gated mechanism as a shared weight to multiply with local and global information to achieve gated global-local linear attention (G2L2Attention). Local information is obtained byLID2Convand we gain global information by introducing 2D Selective Scan (SS2D). These two pathways are integrated through the proposedG2L2Attention, enabling an efficient and consistent fusion of hierarchical spatial features. The extracted features are passed to the decoder. This approach improves road detail representation and provides accurate contextual information. Experiments conducted on three public road datasets demonstrate thatG2L2Netoutperforms existing methods in various evaluation metrics. Our source code is available at: https://github.com/ZehuaChenLab.

Change detection (CD) is a fundamental task in remote sensing imagery, playing a pivotal role in diverse applications such as urban planning, environmental monitoring, and natural disaster assessment. However, most existing methods focus on processing entire images, with limited exploration of cropping images into strip patches and performing operations within these patches. These methods suffer from a waste of computational cost and memory resources in many areas unchanged. In this article, we present a novel paradigm for CD task and introduce a lightweight yet powerful framework, termed SP4CD. Specifically, SP4CD employs a Siamese backbone as a shared feature extractor to capture discriminative and hierarchical representations from bi-temporal remote sensing imagery. Furthermore, an interactive fusion mechanism is devised to integrate the extracted representations across spatial and channel dimensions, thereby enhancing cross-temporal feature interaction. To this end, we design a Patch-wise Intertemporal Channel Integration (PICI) module. By introducing a patch-swapping strategy, PICI integrates dual-temporal feature channels in a nearly parameter-free manner, achieving adaptive channel enhancement while effectively alleviating misclassification errors. In addition, we explore an Intertemporal Synergistic Patch Scanning (ISPS) module. It refines spatial representations through difference-guided selective scan (SS2D) of patches, optimizing detail in changing regions to reduce false negatives. Moreover, two complementary scanning modes (parallel and series modes) are designed based on guidance sequences derived from distinct feature representations, enabling adaptive spatial exploration from multiple perspectives. Extensive experiments conducted on four challenging benchmark datasets verify that the proposed SP4CD achieves competitive or superior performance across multiple evaluation metrics, confirming its effectiveness and generalization capability. Our source code is available at: https://github.com/ZehuaChenLab/SP4CD.

Remote sensing images contain many objects that resemble road structures, making it diffcult to distinguish roads from the background. Moreover, road extraction is affected by many factors, such as lighting conditions, noise, occlusions, etc., resulting in incomplete and discontinuous road extraction. Learning discriminative road features from remote sensing images is a highly challenging task. In this paper, a novel road extraction model is proposed for remote sensing images under encoder and decoder U-Net like architecture. An axial Transformer module (ATM) is designed to learn global road features in the deepest layer with linear computational complexity regarding image size. And a multilayer attention fusion module (MLAF) is also presented to fuse multiple layers of Transformer features, obtaining more comprehensive and richer semantic information. In the skip connection, a channel attention module (CAM) is designed to weight the feature maps along the channel dimension, with the goal of improving the capability of feature representation. Extensive experiments are conducted on the DeepGlobe and Massachusetts road datasets. Compared with other methods, our proposed method in this paper realized road extraction from remote sensing images with higher accuracy and less computational cost, e.g., achieving a intersection over union (IoU) of 81.71% (1.02% improvement) and a 22.38% reduction in convergence time over the latest TransRoadNet on the Massachusetts road dataset. Ablation experiments also demonstrate the effectiveness of the designed model.

Multilevel road feature extraction from remote sensing image plays an important role in numerous applications such as autonomous driving and urban planning. However, interference from background, occlusions, and road-like information makes it difficult to distinguish different levels of roads. To address these issues, this study proposes a deep network named BMDCNet, which adopts LinkNet as the baseline model. The bidirectional multilevel road feature dynamic fusion (BMDF) module is designed to replace the simple skip connections, which greatly reduce the semantic gap between the encoder and the decoder. Furthermore, the dual context dynamic extraction (DCDE) module is designed to dynamically extract and integrate global and local multiscale context information. Finally, experiments are conducted on the DeepGlobe road extraction dataset and the Massachusetts roads dataset. The results demonstrate that compared with LinkNet, F1 and IoU of BMDCNet increased by 1.79% and 2.68% on DeepGlobe, and by 0.35% and 0.47% on Massachusetts, respectively. Our source code is available athttps://github.com/ZehuaChenLab/BMDCNet.

Road extraction from satellite imagery is vital in a broad range of applications. However, extracting complete roads is challenging due to road occlusions caused by the surroundings. This letter proposed an improved encoder–decoder network via extracting road context and integrating full-stage features from satellite imagery, dubbed as RCFSNet. A multiscale context extraction (MSCE) module is designed to enhance inference capabilities by introducing adequate road context. Multiple full-stage feature fusion (FSFF) modules in the skip connection are devised to provide accurate road structure information, and we devise a coordinate dual-attention mechanism (CDAM) to strengthen the representation of road features. Extensive experiments are carried out on two public datasets, and as a result, our RCFSNet outperforms other state-of-the-art methods. The results indicate that the road labels extracted by our method have preferable connectivity. The source code will be available athttps://github.com/CVer-Yang/RCFSNet.

In remote sensing images, roads are usually in complex shapes and can be partially occluded by buildings, trees, and other surroundings. To extract a complete and continuous road network is still a challenging job. This paper proposes a context-aware road extraction method for remote sensing imagery based on Transformer network, in which, a foreground feature enhancement module (FFEM) is designed to further extract detailed road features such as contours from the shallowest feature map; Dual-attention module (DAM) is constructed and applied at different skip connections to make the model focus more on road features in the different level of feature maps; A Swin Transformer-based contextual information extraction module (CIEM) is built between the encoder and decoder modules to capture the global and local road contextual information so as to recover the occluded roads information as much as possible. Furthermore, a multi-scale decoder (M-Decoder) is designed to improve the feature map recovery ability of the decoder module. Experiments on the DeepGlobe road dataset are conducted to verify the efficiency of the proposed method.

Road extraction is a significant research hotspot in the area of remote sensing images. Extracting an accurate road network from remote sensing images is still challenging, because some objects in the images are similar to the road, and some results are discontinuous due to the occlusion. Recently, convolutional neural networks (CNNs) have shown their power in a road extraction process. However, the contextual information cannot be captured effectively by those CNNs. Based on CNNs, combining with high-level semantic features and foreground contextual information (FCI), a novel road extraction method for remote sensing images is proposed in this letter. First, the position attention (PA) mechanism is designed to enhance the expression ability for the road feature. Then, the contextual information extraction module (CIEM) is constructed to capture the road contextual information in the images. At last, an FCI supplement module (FCISM) is proposed to provide foreground context information at different stages of the decoder, which can improve the inference ability for the occluded area. Extensive experiments on the DeepGlobal road dataset showed that the proposed method outperforms the existing methods in accuracy, intersection over union (IoU), precision, and$F1$score and yields competitive recall results, which demonstrated the efficiency of the new model.