FreqSCD: Frequency-Aware Adaptation and Task-Decoupled Learning for SAM2-Based Semantic Change Detection

Semantic change detection in remote sensing imagery plays a pivotal role in urban planning, environmental monitoring, and disaster assessment applications. Existing deep learning-based methods, particularly those relying on triple-branch architectures, often struggle to accurately localize and predict changes in complex spatial environments characterized by diverse land-cover types. To overcome these limitations, this paper proposes a novel network called the Spatial Flow-based Semantic Change Detection Network. This network processes bi-temporal satellite images using a dual-encoder, triple-decoder architecture that progressively refines spatial features at each network stage, improving semantic change detection results. The Attention-Based Siamese Encoder, Cascaded Convolutional Attention Fusion Block, Cascaded Convolutional Attention Refinement Block and Differentiable Binarization layer helps in improving semantic change detection performance. Experimental results of proposed network on the SECOND dataset demonstrate that the proposed model significantly improves the ability to localize critical changes and distinguish between change and no-change regions. The proposed network achieves an overall accuracy of 86.32%, a mean Intersection over Union of 70.33%, a Separated Kappa of 21.21%, and an F1-score for semantic change detection of 66.01%, with a score of 35.94%. These results represent substantial improvements over previous state-of-the-art models, including a 0.26% increase in overall accuracy, a 2.21% increase in mean Intersection over Union, a 2.62% enhancement in Separated Kappa, and a 3.6% improvement in F1-score for semantic change detection compared to the best-performing models. Notably, the proposed network achieves these results with only 14.56 million parameters, making it more effective and efficient than its competitors, which utilize over 22 million parameters.

Change detection (CD) is a crucial task in remote sensing image analysis, including several fundamental tasks such as binary change detection (BCD), semantic change detection (SCD), etc. Recently, deep learning-based models (e.g., Convolutional neural networks (CNN) and Transformers) have made impressive progress in the field of remote sensing CD. However, those CD models are carefully designed for a specific fundamental task (BCD or SCD) and cannot achieve impressive progress in both BCD and SCD tasks simultaneously. In this paper, we designed a generalized architecture for both BCD and SCD tasks by investigating the common characteristics of these two tasks. In the feature extraction stage, a Multi-scale Attention Enhanced Encoder (MSAE) is introduced to extract global context and capture fine-grained features, which is beneficial for both BCD and SCD. In the training stage, a Change Detection Contrastive Learning Module (CDCL) is proposed to optimize feature distribution, improving the discriminative ability in distinguishing change regions and categories. To mitigate class imbalance issues in both BCD and SCD, we introduce Dynamic Rare-Aware Sampling for Change Detection (DRAS-CD), which dynamically prioritizes rare categories and enhances model robustness. Additionally, we collect the Yellow River Basin Semantic Change Detection dataset (YRSCD), which includes 13 change categories across diverse scenes with broad spatial and temporal coverage. Extensive experimental results on four public datasets (WHU-CD, LEVIR-CD, SECOND, Landsat-SCD) and YRSCD have shown superior performance in comparison to other state-of-the-art approaches, especially those based on the visual fundamental model, offering a generalized solution for both BCD and SCD tasks.

Abstract. Semantic change detection (SCD) in remote sensing image aims to identify semantic alterations between bi-temporal images captured at the same geographic location. SCD is extensively applied in fields such as environmental monitoring and disaster assessment. Despite significant advancements in deep learning leading to numerous successful approaches, most existing methods primarily rely on visual representation learning, thereby overlooking the potential benefits of multimodal data. Recently, vision-language models have demonstrated outstanding performance across various downstream tasks. In this paper, we propose a novel framework named TextSCD that leverages text-based semantic information to guide the generation of semantic change maps. Our approach integrates Gemini to generate change descriptions between bi-temporal images and employs a multi-level semantic extraction method to capture features from both images and their corresponding captions. Furthermore, we introduce a semantic text-guided interaction module that facilitates the effective integration of visual and textual features, enhancing multimodal knowledge transfer and the extraction of discriminative features. This design effectively reduces false detections and omissions. We validate the effectiveness of our model on the SECOND dataset, achieving notable improvements in overall accuracy for semantic change detection.

Semantic change detection (CD) not only helps pinpoint the locations where changes occur, but also identifies the specific types of changes in land cover and land use. Currently, the mainstream approach for semantic CD (SCD) decomposes the task into semantic segmentation (SS) and CD tasks. Although these methods have achieved good results, they do not consider the incentive effect of task correlation on the entire model. Given this issue, this article further elucidates the SCD task through the lens of multitask learning theory and proposes a semantic change detection network based on boundary detection and task interaction (BT-SCD). In BT-SCD, the boundary detection (BD) task is introduced to enhance the correlation between the SS task and the CD task in SCD, thereby promoting positive reinforcement between SS and CD tasks. Furthermore, to enhance the communication of information between the SS and CD tasks, the pixel-level interaction strategy and the logit-level interaction strategy are proposed. Finally, to fully capture the temporal change information of the bitemporal features and eliminate their temporal dependency, a bidirectional change feature extraction module is proposed. Extensive experimental results on three commonly used datasets and a nonagriculturalization dataset (NAFZ) show that our BT-SCD achieves state-of-the-art performance. The code is available at https://github.com/TangYJ1229/BT-SCD.

Abstract. Semantic change detection (SCD) focuses on identifying changes in surface coverage while simultaneously classifying the types of changes. This approach provides detailed information valuable for urban planning, environmental monitoring, and other applications, making it a key area of interest in remote sensing research. Despite recent advances, existing SCD studies are hindered by the lack of high-resolution satellite imagery datasets and insufficiently comprehensive semantic label coverage in publicly available datasets. To address these limitations, we have developed a large-scale high-resolution remote sensing dataset consisting of 11,587 satellite image pairs, each with 1-meter spatial resolution and a size of 512 × 512 pixels, representing land cover changes across Beijing between 2017 and 2018. This dataset encompasses diverse land surface scenes with comprehensive semantic annotations. Furthermore, it includes full-coverage semantic segmentation labels from pre-change phases and a larger sample size of 2048 × 2048 pixels to support future research on multi-class and large-format change detection. We benchmark eight state-of-the-art SCD algorithms using this dataset, providing critical performance metrics that serve as valuable references for subsequent research. This dataset not only addresses existing gaps but also establishes a robust foundation for advancing deep learning-based semantic change detection, enabling more accurate and comprehensive analysis of complex and diverse land cover changes. More information about the project can be found at https://github.com/17x-osborn/HRMS-SCD.

Semantic feature-constrained multitask siamese network for building change detection in high-spatial-resolution remote sensing imagery

Semantic change detection (SCD) aims to recognize land cover transitions from remote sensing images of the given scene acquired at different times. The semantic change maps produced by SCD can provide not only the locations of changes but also the detailed change types (e.g., “from-to” change type). This exhaustive change information plays a significant role in various applications. Postclassification methods with multitemporal remote sensing images have been widely used in SCD. However, many existing methods suffer from the accumulation of misclassification errors. In this article, a deep Siamese postclassification fusion network (PCFN) is proposed to address this problem. PCFN is composed of the Siamese classification network (SCN) for land cover mapping (LCM) and soft fusion network (SFN) for postclassification SCD, respectively. In PCFN, SCN is designed to effectively integrate the temporal correlation between two images by processing the joint features, further improving classification performance. Then, SFN is constructed to determine changes and identify the specific change types through automatic soft fusion. SFN fuses the multitemporal LCM features generated by SCN, then adaptively maps them to the decision space for soft fusion during network training, and predicts the final semantic change maps. Extensive experimental results on two challenging SCD datasets demonstrate that our method can alleviate the error accumulation effectively by combining temporal correlation integration and soft fusion, and achieve promising performance superior to the other state-of-the-art methods in SCD.

The rapid advancement in Earth observation technologies has improved the acquisition of precise data on terrestrial changes. However, traditional binary change detection (BCD) fails to satisfy the complex demands of contemporary applications. In contrast, semantic change detection (SCD) can identify changed areas and further detect the types of changes in ground objects. Nonetheless, current SCD approaches pose significant challenges, including the inadequate extraction and utilization of multi-level features and limited synergistic integration of related tasks. This paper proposes a multi-level feature-aggregated network (MLFA-Net), which decomposes the SCD task into two related subtasks: ground object classification and BCD of bi-temporal images. To optimize the classification performance, MLFA-Net incorporates feature alignment and cross-level feature aggregation modules that operate between the encoders and decoders, achieving effective extraction, aggregation, and utilization of multi-level image features. MLFA-Net also integrates symmetric transform feature aggregation modules between classification and BCD branches, leveraging prior knowledge to facilitate cross-task transmission of multi-level information and further optimize SCD results. MLFA-Net achieved Scores of 0.6214 and 0.3581 on Landsat-SCD and SECOND datasets, respectively, with improvements of at least 0.0402 and 0.0035 compared to other methods. Quantitative and qualitative experiments validate the efficacy and superiority of the proposed MLFA-Net in SCD tasks.

Semantic Change Detection (SCD) in remote sensing imagery is a common technique for monitoring surface dynamics. However, geospatial data acquisition increasingly involves the collection of visible and infrared images. SCD in visible and infrared image pairs confronts the challenge of distinguishing genuine semantic change from spectral discrepancies caused by heterogeneous imaging mechanisms. To address this issue, we propose a Modal Feature Analysis Semantic Change Detection Network (MFA-SCDNet), a novel framework that analyzes cross-modal features for change identification. The proposed architecture operates through three principal technical components: An infrared feature enhancement module that transforms infrared inputs into three-channel representations through spectral domain adaptation, enhancing the network’s perception of both high-frequency and low-frequency information in images; an encoder–decoder structure that simultaneously extracts modality-specific features and common features through adversarial learning; and a synergistic information fusion mechanism that integrates semantic recognition with change detection through multi-task optimization. Specific features are employed for semantic recognition, while common features are utilized for change detection, ultimately resulting in a comprehensive understanding of semantic changes. Experiments on public datasets show that MFA-SCDNet has an average improvement of 9.4% in mIoUbc and 12.9% in mIoUsc compared with the alternatives. MFA-SCDNet has better performance in heterogeneous images SCD.

Semantic change detection (SCD) aims to identify land cover changes between bi-temporal remote sensing images and plays a crucial role in applications, such as urban monitoring and disaster assessment. Most existing methods rely on a shared-weight encoder with a dedicated change extractor, which limits their ability to precisely localize changed regions and maintain intra-class semantic consistency. Although vision foundation models (VFMs) exhibit strong generalization ability, their potential for SCD remains largely underexplored. In this paper, we introduce ChangeVFM, a novel yet effective framework that unleashes the power of vision foundation models for semantic change detection in remote sensing images. This framework is supplemented by a spatiotemporal modeling module to capture fine-grained spatial details and a query-based feature injector that integrates VFM’s semantic priors with multi-scale spatiotemporal features. The feature injector ensures that the ChangeVFM excels in both maintaining semantic consistency and multi-scale information of bi-temporal images. Without bells and whistles, ChangeVFM achieves competitive performance on the HRSCD, SECOND, and Landsat-SCD. Comprehensive quantitative and qualitative experiments further validate the effectiveness of the introduced modules and the robustness of the proposed method.

Semantic Change Detection (SCD) in remote sensing imagery requires models that integrate extensive spatial context for broad geographic patterns, computational efficiency for large-scale datasets, and sensitivity to class-imbalanced land-cover transitions to detect rare or asymmetric changes. Early SCD approaches relied on Convolutional Neural Networks, which excel in local feature extraction but falter in modeling global spatial context due to limited receptive fields. Transformers mitigate this by capturing long-range dependencies via self-attention, yet their quadratic complexity impairs efficiency on vast remote sensing data. Emerging Mamba architectures, based on state-space models, strike a balance with linear complexity and robust long-range modeling, delivering efficient global context capture and improved performance. In this study, we introduce Mamba-FCS, an SCD framework leveraging a Visual State Space Model backbone, with three key contributions: (1) a Joint Spatio-Frequency Fusion block that integrates log-amplitude frequency-domain features to sharpen edges and mitigate illumination artifacts, (2) a Change-Guided Attention (CGA) module that explicitly bridges the intertwined Binary Change Detection and SCD tasks, and (3) a novel loss function inspired by the Separated Kappa (SeK) metric to optimize for class imbalance. Experiments on the benchmark datasets show that Mamba-FCS consistently outperforms recent state-of-the-art algorithms. Ablation studies indicate that spatio–frequency fusion and CGA mainly sharpen boundaries and suppress hallucinated changes, while the SeK-inspired loss improves minority-class semantics. These results highlight the potential of Mamba-FCS as a scalable and accurate approach for remote sensing change detection. Source code and configuration files are available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/Buddhi19/Mamba-FCS.git</uri>

Semantic change detection (SCD) is a newly important topic in the field of remote sensing (RS) image interpretation since it provides semantic comprehension for bi-temporal RS images via predicting change regions and change types and has great significance for urban planning and ecological monitoring. With the availability of large scale bi-temporal RS datasets, various models based on deep learning (DL) have been widely applied in SCD. Since convolution operators in DL extracts two-dimensional feature matrices in the spatial dimension of images and stack feature matrices in the dimension termed the channel, feature maps of images are tri-dimensional. However, recent SCD models usually overlook the stereoscopic property of feature maps. Firstly, recent SCD models are usually limited in capturing spatial global features in the process of bi-temporal global feature extraction and overlook the global channel features. Meanwhile, recent SCD models only focus on spatial cross-temporal interaction in the process of change feature perception and ignore the channel interaction. Thus, to address above two challenges, a novel fine-grained feature perception network (FFPNet) is proposed in this paper, which employs the Omni Transformer (OiT) module to capture bi-temporal channel–spatial global features before utilizing the Omni Cross-Perception (OCP) module to achieve channel–spatial interaction between cross-temporal features. According to the experiments on the SECOND dataset and the LandsatSCD dataset, our FFPNet reaches competitive performance on both countryside and urban scenes compared with recent typical SCD models.

Semantic change detection (SCD) in remote sensing images can simultaneously obtain changed areas and the transformation of ground objects, providing finer-grained information support in ground observation applications. Recently, multi-task networks have become the main paradigm for SCD, but their detection performance is still affected by limited multi-scale feature extraction and insufficient cross-task information interaction. To address these issues, we propose a global-local awareness integrated network (GLAI-Net) for SCD in remote sensing images. We design a parallel encoding structure and utilize convolutional neural networks and Transformer to achieve multi-scale modeling of images and enhance feature expression ability. Meanwhile, we propose multi-scale feature fusion (MSFF) modules in GLAI-Net to enhance the focus of detail features on changed objects with different sizes. Between the classification and change detection decoding branches, we propose semantic change response (SCR) modules in GLAI-Net that fully utilize the correlation between different tasks to improve the consistency and accuracy of detection results. Qualitative and quantitative results on the Landsat-SCD and SECOND datasets indicate that the proposed GLAI-Net outperforms the comparison methods in SCD performance, with SeKs of 0.5182 and 0.2063, respectively. In addition, the ablation experiment results confirm the effectiveness of MSFF and SCR modules in improving feature diversity and optimizing SCD performance.

Semantic change detection (SCD) involves the simultaneous extraction of changed regions and their corresponding semantic classifications (pre- and post-change) in remote sensing images (RSIs). Despite recent advancements in vision foundation models (VFMs), the fast-segment anything model has demonstrated insufficient performance in SCD. In this article, we propose a novel VFMs architecture for SCD, designated as VFM-ReSCD. This architecture integrates a side adapter (SA) into the VFM-ReSCD to fine-tune the fast segment anything model (FastSAM) network, enabling zero-shot transfer to novel image distributions and tasks. This enhancement facilitates the extraction of spatial features from very high-resolution (VHR) RSIs. Moreover, we introduce a recurrent neural network (RNN) to model semantic correlation and capture feature changes. We evaluated the proposed methodology on two benchmark datasets. Extensive experiments show that our method achieves state-of-the-art (SOTA) performances over existing approaches and outperforms other CNN-based methods on two RSI datasets.

Recent change detection (CD) methods focus on the extraction of deep change semantic features. However, existing methods overlook the fine-grained features and have the poor ability to capture long-range space–time information, which leads to the micro changes missing and the edges of change types smoothing. In this paper, a potential transformer-based semantic change detection (SCD) model, Pyramid-SCDFormer is proposed, which precisely recognizes the small changes and fine edges details of the changes. The SCD model selectively merges different semantic tokens in multi-head self-attention block to obtain multiscale features, which is crucial for extraction information of remote sensing images (RSIs) with multiple changes from different scales. Moreover, we create a well-annotated SCD dataset, Landsat-SCD with unprecedented time series and change types in complex scenarios. Comparing with three Convolutional Neural Network-based, one attention-based, and two transformer-based networks, experimental results demonstrate that the Pyramid-SCDFormer stably outperforms the existing state-of-the-art CD models and obtains an improvement in MIoU/F1 of 1.11/0.76%, 0.57/0.50%, and 8.75/8.59% on the LEVIR-CD, WHU_CD, and Landsat-SCD dataset respectively. For change classes proportion less than 1%, the proposed model improves the MIoU by 7.17–19.53% on Landsat-SCD dataset. The recognition performance for small-scale and fine edges of change types has greatly improved.