Search

Against the backdrop of urbanization and land use policy formulation, the accurate identification of changes is imperative for the effective monitoring and control of urban growth, conservation of natural resources, and environmental protection. Traditional remote sensing techniques for change detection, foundational to these applications, typically face significant challenges such as pronounced errors and inadequate parsing of complex environments. To surmount these obstacles, this study proposes a cutting-edge, deep learning-based strategy with a Siamese network framework, and utilizing a modified U-Net++ network (SIAM-CDNET) to improve change detection capabilities. This model integrates an attention mechanism to enhance the acuity and distinguishing features of remote sensing imagery, particularly for subtle variations and intricate structures in urban settings. Furthermore, an edge-detection component designed for multiscale feature amalgamation and a change detection-specific composite loss function are devised to increase the precision and generalizability of the model. Evaluations based on the LEVIR-CD,CDD and GZ-CD datasets indicate that the proposed SIAM-CDNET markedly surpasses existing methods in terms of accuracy and F1-score, with values of 90.50% and 90.82% (LEVIR-CD dataset),89.95% and 91.18% (CDD dataset) and 93.81% and 90.8%(GZ-CD dataset), respectively. These results confirm the exceptional performance of SIAM-CDNET in urban development and land use change detection, potentially making a significant technical contributions to the field.

Remote sensing image change detection is a valuable technology for analyzing the earth observation data. It has significant application value in resource monitoring, disaster assessment, and urban planning. However, current change detection methods have not fully explored the interrelationships between bitemporal data, and the extraction process of change information lacks prior guidance and constraints. Therefore, it is easy to produce missed detections and false alarms when facing complex backgrounds and variable objects in remote sensing images. To tackle such issues, we propose a cross-scale guided high-order feature interaction change detection network for dual temporal images. Specifically, a cross-scale guided dual encoder–decoder backbone is proposed to constrain the reconstruction process of change objectives, and guide geometric prior to optimize the representation of target structures. Next, an efficient high-order feature interaction module is designed, employing multilevel receptive fields to enhance the perception ability for multiscale features. Moreover, we construct a bitemporal feature alignment fusion module, which decouples and filters out the interference of background pseudo changes through interactive perception of spatial–temporal differences. Comprehensive experimental validation is undertaken on four representative change detection datasets (LEVIR-CD, WHU-CD, DSIFN-CD, and S2Looking). The findings demonstrate that the network demonstrated state-of-the-art performance.

Remote sensing image change detection aims to extract high-level semantic feature to identify the changed areas (CAs) between dual-temporal images (DTIs). However, the diversity in the CA shape and size poses certain challenge to the change detection (CD) task. Besides, different illumination conditions in the same scene of the DTI further increase the CD difficulty. In response to these above issues, this article proposes a multiscale feature fusion CD network-WSMsFNet, which fully utilizes the local and global information of multiscale features to achieve comprehensive representation of the change scene. In addition, the network improves the feature extraction ability of each module through the whole process supervision loss function. First, the network hierarchically extracts different scale information of the two temporal RS. Then, special information enhancement and fusion modules are constructed for various feature layers (i.e., the same level, adjacent level, and global features), aiming to enhance the local feature representation ability and contextual information relevance of the deep network. Finally, the whole-process loss function is set to supervise the intermediate layer learning, which can effectively enhance the feature representation ability and guide feature extraction direction of each module. Experiments have shown that the WSMsFNet has achieved significant results in both qualitative and quantitative indicators.

Forest fires are significant disturbances to ecosystems, necessitating accurate mapping of burned areas and assessment of burn severity. First, we reconstruct a dataset whose label uses a more flexible classification method from Landsat imagery and establish auxiliary environmental datasets for fire-affected regions. Leveraging vegetation change prefire and postfire, we propose a transformer-based change detection model that integrates remote sensing and environmental information effectively. We introduce a multilevel feature fusion mechanism to address spatial resolution degradation in burn severity estimation. Experimental results show our model closely approximates evaluation dataset labels. For burned area segmentation, our method achieves the highest F1 (0.897) and mIoU of 0.781. For burn severity estimation, our method also achieves the highest mIoU (0.851). Incorporating auxiliary features improves performance by nearly 30%, while the multilevel feature fusion mechanism reduces resolution degradation by 9.6%.

Snow avalanches constitute a dangerous natural hazard in snow-clad mountain regions. Spaceborne synthetic aperture radar (SAR) has emerged as an effective tool for monitoring avalanche activity also in remote areas due to its all-weather capabilities and sensitivity to the presence of avalanche debris. The objective of this study was the application of a novel polarimetric change detection approach to the task of identifying avalanche debris candidate areas and characterizing the impact of factors, such as avalanche size and topography, on its performance. We applied polarimetric change vectors (PCVs) for change detection between pairs of Sentinel-1 SAR images acquired over the Swiss Alps during a two-month period characterized by episodes of exceptionally high avalanche activity. PCV were classified using a two-step mixture analysis of their magnitude and direction components. Candidates for avalanche debris were matched against a reference database of $>$ 16 000 avalanche outlines. The detected changes were largely attributed to snow processes, such as transitions from dry to wet snow and vice versa, as well as the occurrence of avalanches. 23% of all reference avalanche polygons could be matched to retrieved avalanche debris patches. The matching rate strongly depended on avalanche size, reaching $>$ 33% for very large (size 4) avalanches and, to a lesser degree, on the orientation of the avalanche relative to the look direction of the sensor. Apart from constituting a standalone approach for detecting avalanche debris candidate areas, the PCV method has potential for integration into automatic avalanche detection workflows based on machine learning methods.

Semantic change detection (SCD) extends the traditional change detection (CD) task to simultaneously identify the change areas and their corresponding land cover categories in bi-temporal images. This “from-to” change information holds significant value in numerous practical applications and is increasingly garnering attention in the remote sensing domain. However, prevalent challenges such as loss of detail and class imbalance significantly hinder the efficacy of SCD applications. To address this challenge, we propose a novel semantic information collaboration network (SIC-Net). This network incorporates a detail capture path and a spatial-temporal semantic coordination module aiming to effectively execute SCD by fusing detailed information with contextual features and harnessing synergies between binary CD and semantic segmentation. Additionally, we introduce a pseudo-label growth algorithm to mitigate the substantial loss of sample category information. The experimental results on two widely used SCD datasets demonstrate that SIC-Net outperforms other methods across various evaluation metrics, achieving SeK of 23.96% and 61.29%, respectively. These findings not only validate the effectiveness of the SIC-Net but also provide new ideas and directions for future research in the field of remote sensing, particularly in SCD.

Foundation models have reshaped the landscape of remote sensing (RS) by enhancing various image interpretation tasks. Pretraining is an active research topic, encompassing supervised and self-supervised learning methods to initialize model weights effectively. However, transferring the pretrained models to downstream tasks may encounter task discrepancy due to their formulation of pretraining as image classification or object discrimination tasks. In this study, we explore the multitask pretraining (MTP) paradigm for RS foundation models to address this issue. Using a shared encoder and task-specific decoder architecture, we conduct multitask supervised pretraining on the segment anything model annotated remote sensing segmentation dataset, encompassing semantic segmentation, instance segmentation, and rotated object detection. MTP supports both convolutional neural networks and vision transformer foundation models with over 300 million parameters. The pretrained models are finetuned on various RS downstream tasks, such as scene classification, horizontal, and rotated object detection, semantic segmentation, and change detection. Extensive experiments across 14 datasets demonstrate the superiority of our models over existing ones of similar size and their competitive performance compared to larger state-of-the-art models, thus validating the effectiveness of MTP.

Change detection is an important method of analyzing information about changes in geographical features. However, existing deep learning feature difference methods often lead to the loss of detailed information. Differences in features can arise from factors like illumination or geometric variations rather than actual change regions, resulting in inaccurate change detection. This leads to poor detection of fine-grained boundaries and internal hole problems. To alleviate this, we propose a novel change detection network guided by change boundary awareness and incorporating the concept of boundary-to-center. Our network introduces a change boundary-aware module to capture boundary information of change regions. This module enhances boundaries, reducing the influence of noise in feature differences and providing rich contextual information to improve the accuracy of change boundaries. Additionally, we propose a bitemporal feature aggregation module (BFAM) based on spatial-temporal features. The BFAM aggregates multiple receptive fields features and complements texture information. Both modules utilize the SimAM attention mechanism to enhance the finegrained nature of the features. In addition, we introduce a deep feature extraction module to extract deep features and minimize information loss during the decoupling process. The proposed change detection network in this article is guided by change boundary perception, progressively integrating semantic and spatial texture information to refine edges and enhance internal integrity. The performance and efficiency of B2CNet have been validated on four publicly available remote sensing image change detection datasets. Through extensive experiments, the effectiveness of the proposed method has been demonstrated. For example, in terms of IOU for LEVIR, WHU, SYSU, and HRCUS datasets, the improvements compared to the baseline are 1.89%, 2.86%, 4.70%, and 3.79%, respectively.

Remote sensing image change detection, as one of the important branches of remote sensing technology, has been greatly developed and improved in recent years. The effectiveness of change detection is unstable, as deep learning components exhibit a high degree of sensitivity to hyperparameters, necessitating substantial fine-tuning to achieve optimal outcomes. Additionally, large and intricate deep learning models are not appropriate for smaller-scale datasets. In this paper, a remote sensing image change detection method based on general deep forest module (GDFM) is proposed. In the proposed GDFM, the common remote sensing image change detection method is firstly used for preliminary detection to obtain the detection result to be optimized, and the optimized change detection result is then obtained by using the classification function of the deep forest. This work aims to achieve a general enhancement of the performance of existing change detection methods by utilizing the characteristics of multi-grained scanning in deep forest and the data classification capabilities of cascaded forests. At the same time, this model alleviates the problems of parameter tuning complexity and inapplicability to small datasets in additional deep learning modules. It is worth noting that this general method is not a direct splicing process, but combines the advantages of the two parts and uses the characteristics of the deep forest to achieve the goal. Experiments are conducted on multiple common remote sensing change detection models. The results show that the proposed GDFM has significant improvements in metrics, which include varying degrees of improvement in F1 score by around 1% -14%.

Synthetic aperture radar (SAR) image change detection is currently a popular topic, but the existence of speckle noise renders it challenging. Making full use of the neighborhood information can reduce noise interference and improve accuracy. To this end, this study proposes a graph-guided method for SAR image change detection. First, we establish the local, nonlocal, and global connections according to three types of neighboring rules. A novel heterogeneous graph is then constructed by assigning different weights to these multilevel connections. On the basis of heterogeneous graph, a composite random walk matrix (CRWM) is presented to quantify the similarity of multilevel connections. Thereafter, the heterogeneous graph shift system consisting of the multiorder CRWM is designed to aggregate the attributes of neighboring vertices along the multiorder and multilevel connections. The change measure can be implemented by comparing the output signals from the heterogeneous graph shift systems, resulting in the generation of difference image with good separability. Finally, change analysis is carried out using a binary classification algorithm. Experiments conducted on six real SAR datasets confirm that the proposed M2HG method successfully strikes a balance between speckle suppression and change enhancement. This strategic balance translates into improvements in performance metrics, with OA, κ, and F1 indicators surpassing those of the suboptimal method by 0.87%, 3.34%, and 3.84%, respectively. Overall, the proposed method emerges as a promising contender for SAR image change detection tasks.

The scattering properties of targets in polarimetric synthetic aperture radar (PolSAR) images are directly influenced by the targets' orientations, as the scattering properties from the same target with different orientations can be very different. This interpretation diversity caused by the target orientations is one of the primary technical bottlenecks in PolSAR image interpretation. In this article, a 3-D convolutional vision transformer (3-D-Conv-ViT) is proposed to describe the relationship between polarimetric coherent matrices with different polarization orientation angles (POAs) for PolSAR image classification and change detection. First, 3-D convolutional neural networks are used to capture the high-level feature representations of the polarimetric coherent matrix sequence. Second, a new Rotation-3-D-ViT block is proposed to learn the local and global representations of the high-level feature maps. The self-attention mechanism in the ViT can express the regularity of polarimetric coherent matrices with different POAs and improve the PolSAR image interpretation performance. Third, combined with different classifiers, the proposed 3-D-Conv-ViT can be applied to both PolSAR image classification and change detection. Experiments on real PolSAR image datasets demonstrate that the proposed method can overcome the problem of the interpretation ambiguity caused by the target orientation. The classification accuracies of the proposed method can reach 94.01%–99.48%, and the change detection accuracies can reach 93.84%–96.86%.

Urban topographic changes have a crucial impact on the environment, which serves as the foundation for human survival. However, due to the different spectral response characteristics of various remote sensing satellite sensors and the presence of aging issues, the captured remote sensing images exhibit color imbalance under different seasons and lighting conditions. Existing change detection methods fail to establish the correlation between changed and unchanged information in color-imbalanced urban topographic remote sensing images, thereby reducing the utilization of feature extraction information. Additionally, using fixed thresholds for region determination may inaccurately and incompletely generate change areas. Therefore, this paper proposes a remote sensing image change detection method based on multi-level and multi-diversity feature fusion. Firstly, a ResNet-18 network is employed to extract dual-temporal features. Secondly, to enhance the discrimination between change and unchanged regions during feature extraction and improve the utilization of feature extraction information, a Diversity Feature Fusion Module (DFFM) is designed to reduce false alarms and omissions in the detection results. Furthermore, to effectively address the problem of boundary misjudgment in change areas caused by fixed thresholds, an Adaptive Threshold Module is devised to adaptively learn the texture features of change and unchanged regions, enabling the generation of more accurate thresholds for boundary determination, thereby improving the robustness of the algorithm model and alleviating false alarms. Finally, experimental tests demonstrate that our method achieves excellent performance in two public change detection datasets.

Change detection of remote sensing images is an essential method for observing changes on the Earth's surface. Deep learning can efficiently process remote sensing images. However, shallow features in remote sensing data from different time are inherently inconsistent. During the feature extraction stage, these shallow features are mapped onto different dimensional feature maps, giving rise to noise information. Existing algorithms are ineffective in dealing with noise effectively. This can lead to detection results being influenced by shallow features noise information, resulting in fake detections. To address this issue, feature filtering fusing network (F3Net) is proposed in this article. In F3Net, feature filtering and aggregation (FFA) module is designed to integrate bitemporal remote sensing features, which initially filters out noise information from different temporal domains. In addition, the channel feature difference fusion (CFDF) module is introduced to fuse high-dimensional features. Within CFDF, channel information filtering convolution is utilized to filter out noise information from high-dimensional feature channels across multiple receptive fields. In order to verify the performance of F3Net, comparative experiments were conducted on multiple public datasets with other state-of-the-art models, and F3Net achieved the best performance.

Over the last decade, there has been an increasing frequency and intensity of wildfires across the globe, posing significant threats to human and animal lives, ecosystems, and socioeconomic stability. Therefore, urgent action is required to mitigate their devastating impact and safeguard earth's natural resources. Robust machine learning methods combined with the abundance of high-resolution satellite imagery can provide accurate and timely mappings of the affected area in order to assess the scale of the event, identify the impacted assets, and prioritize and allocate resources effectively for the proper restoration of the damaged region. In this work, we create and introduce a machine-learning-ready dataset we name FLOGA (Forest wiLdfire Observations for the Greek Area). This dataset is unique as it comprises satellite imagery acquired before and after a wildfire event with variable spatial and spectral resolution and contains a large number of events, where the corresponding burnt area ground truth has been annotated by domain experts. FLOGA covers the wider region of Greece, which is characterized by a Mediterranean landscape and climatic conditions. We use FLOGA to provide a thorough comparison of specialized spectral indices as well as multiple machine learning and deep learning algorithms for the automatic extraction of burnt areas, approached as a change detection task. Finally, we propose a novel deep learning model, namely BAM-CD. Our benchmark results demonstrate the efficacy of the proposed technique in the automatic extraction of burnt areas, outperforming all other methods in terms of accuracy and robustness.

Infrared small target detection refers to the extraction of small targets in a complex, low signal-to-noise ratio background. Depthwise convolution makes it difficult to comprehensively characterize small infrared targets and ignores the importance of image background for the detection task. In this article, we propose the generative adversarial differential analysis (GADA) model for infrared small target detection, the core of which aims to weaken the reliance on target features and enhance the use of background information. Specifically, we first construct pseudobackground labels by the fast marching method. Then, the background-guided generative adversarial network is used to learn the background data distribution. On this basis, the differential image containing interest regions of small targets is obtained by differential analysis. Finally, the detection results are obtained by performing an elaborate characterization of the interest regions. The effectiveness of GADA is verified with three public datasets. Compared to several state-of-the-art methods, GADA achieves better performance in terms of $F1$, $IoU$, and $AUC$.

Hyperspectral images are valuable for precise land cover change detection in a consistent area over time. Nevertheless, supervised methods for hyperspectral change detection face limitations due to insufficient labeled samples. Additionally, current deep learning-based approaches often neglect critical feature interactions across temporal, spatial, and spectral domains. To address these challenges, we introduce a semisupervised model called the HyperMatch-based Cross Temporal and Spatial Interaction Transformer (CTSIT) for hyperspectral change detection. The key contributions of this study are as follows: Introduction of the HyperMatch training schedule, which relies on weak-to-strong consistency, to enhance feature extraction using both labeled and unlabeled data. Then, the contribution of the Cross Temporal Bidirectional Attention module to emphasize bidirectional temporal interactions and resemblances. Finally, Introduction of the Spatial and Spectral Attention module, which includes the Dense Cross Spatial Attention and Spectral Attention modules, to capture long-range densely spatial interactions and internal spectral similarities. Extensive comparative experiments conducted on three mainstream hyperspectral change detection datasets confirm the effectiveness and superiority of the proposed HyperMatch-based CTSIT in utilizing both labeled and unlabeled samples. For instance, the training ablation demonstrates that this method significantly outperforms most existing state-of-the-art hyperspectral change detection methods, even with a significantly smaller number of samples.

High-resolution stereo satellite images (HRSSIs) have the potential to provide the accurate height and volume information, playing a crucial role in assessing building collapses during various natural disasters. However, the time-consuming process of three-dimensional (3-D) reconstruction, inadequate vertical accuracy of digital surface model (DSM), and concentrated clustering of buildings pose challenges for collapse assessment focused on buildings. Therefore, we present an improved approach for rapid fine-grained assessment of building collapses. First, the accurate and consistent positioning parameters for HRSSIs are obtained through the combined block adjustment using laser altimetry points, ensuring the generation of DSMs with vertical accuracy exceeding 2 m. Next, a set of rapid 3-D reconstruction techniques is introduced, achieving a significant eightfold improvement in generating DSMs. Subsequently, we deploy an automated workflow for batch processing and registration of open-source building footprints, enabling the accurate extraction of building height changes from dual-time DSMs. Finally, based on the building change image, a large-scale GIS image of building floor-level collapses is generated using connected component detection and threshold classification strategies. These findings have far-reaching implications for postdisaster emergency response, damage assessment, and expeditious reconstruction efforts. In our study, we processed an 800 km2 area in Kahramanmaras Province, Turkey, generating dual-time DSMs within 1 h. This enabled the assessment of floor-level collapses for a total of 48 092 buildings within the area. Validation was conducted on 361 houses in the city center, utilizing Google Street view images as ground truth. Remarkably, our approach achieved a high accuracy rate of 93.27% in floor-level assessment.

Accurate change detection using remote sensing data is crucial for understanding surface dynamics. Despite the impressive success of current convolutional neural network (CNN)-based techniques, their feature extraction and representation capabilities are limited, leading to pseudochanges and omissions. To address this issue, we propose a multiscale fusion CNN-transformer network (MSFCTNet), which incorporates the strengths of CNN and transformer to improve change detection capability. The network utilizes a Siamese CNN to extract features from the bi-temporal image pairs and then combines multiscale features using a CNN-transformer hybrid structure to extract global and local features. In the decode stage, a gated attention module is used to filter the extracted features layer by layer. Moreover, before outputting the prediction results, a feature refinement head is employed to further refine the features, suppress background noise, and improve detection capability. LEVIR-CD, SVCD, and SYSU-CD datasets are used to evaluate the proposed network. Experimental results indicate that MSFCTNet outperforms other state-of-the-art change detection methods, proving the potential of MSFCTNet for change detection tasks.

Remote sensing image change detection plays an important role in urban planning and environmental monitoring. However, the existing change detection algorithms have limited ability in feature extraction, feature relationship understanding, and capture of small target features and edge detail features, which leads to the loss of some edge detail information and small target features. To this end, a new dual-attention-guided multiscale feature aggregation network is proposed. In the encoding stage, the fully convolutional dual-branch structure is used to extract the semantic features of different scales, and then, the multiscale adjacent semantic information aggregation module is used to aggregate the adjacent semantic features at different scales, which can better capture and fuse the features of different scales, thereby improving the accuracy and robustness of change detection. In the decoding stage, the dual-attention fusion module is proposed to guide and fuse the features extracted from different scales along the spatial and channel directions and reduce the background noise interference. In addition, this article also proposes a three-branch feature fusion module and a global semantic information enhancement module to make the network better integrate global semantics and differential semantics and further integrate high-level semantic features. We also introduce an auxiliary classifier in the decoding stage to provide additional supervision signals and fuse the output of the three auxiliary classifiers with the output of the main decoder to further achieve multiscale feature fusion. The comparative experiments on three remote sensing datasets show that the proposed method is superior to the existing change detection methods.

In this study, we examine the potential of leveraging self-supervised learning (SSL) and transfer learning methodologies for forest disturbance mapping using Earth Observation (EO) data. Our focus is on natural disturbances caused by windthrow and snowload damages. Particularly, we investigate the potential of knowledge-distillation-based contrastive learning approaches to obtain comprehensive representations of forest structure changes using Copernicus Sentinel-1 and Sentinel-2 satellite imagery. Leveraging pretrained backbone models from knowledge distillation, we employ transfer learning based on deep change vector analysis to delineate forest changes. We demonstrate developed approaches on two use cases, namely, mapping windthown forest using bitemporal Sentinel-1 and Sentinel-2 images, and mapping forest areas damaged by excessive snowload using bitemporal Sentinel-1 images. Developed self-supervised models were compared in a benchmarking exercise. The best results were provided by pixel-level contrastive learning for Sentinel-1-based snowload damage mapping with an overall accuracy of 84% and an $F_{1}$ score of 0.567, and for Sentinel-2-based forest windthrow mapping with an overall accuracy of 76.5% and an $F_{1}$ score of 0.692. We expect that developed methodologies can be useful for mapping also other types of forest disturbances using Copernicus Sentinel images or similar EO data. Our findings underscore the potential of SSL and transfer learning for enhancing forest disturbance detection using EO.

In recent years, detecting anomalies in networks has attracted much attention in the literature. Researchers have developed various statistical process control (SPC) techniques to monitor the communication level within the networks. It is quite common to see the nodes belonging to some densely connected communities or subgroups and the nodes within a community communicate with each other more frequently than with nodes from other communities. However, we can find scarce SPC techniques that consider the multiple community labels and heterogeneity of nodes simultaneously. It is necessary to overcome this drawback to improve the change detection efficiency. This article establishes a new mixed-membership stochastic block model to characterize the multiple community labels and heterogeneity of nodes simultaneously. Then a new control chart is developed based on generalized likelihood ratio test (GLRT). The proposed control chart can be easily implemented in practice and avoids constant parameter estimation in Phase II monitoring. Numerous simulation studies and a real example show the advantages of proposed chart over existing ones.

The integration of the transformer and convolutional neural network (CNN) has become a useful method for change detection in remote sensing images. The main function of the transformer is to capture the global features, while the CNN is more for obtaining the local features. However, such an integration is not efficient for change detection in the very-high-resolution (VHR) remote sensing images with fine surface detail information. Hence, to improve this traditional construction of the transformer and CNN, we propose a dense Swin-Transformer-V2 (DST) and VGG16, coined as DST-VGG, for extracting the discriminatory features for change detection. The difference between our proposed network and other networks is that the output of the VGG16 encoders will be used in the DST in which more Swin-V2 blocks are added for fine feature extraction. The learning model in the VGG16 encoders employs a self-supervised method, which is guided through the change in details. Our network not only inherits the advantages of the integration of the transformer and CNN, but also captures the features of change relationship through the DST and catches the primitive features in both prechanged and postchanged regions through the VGG16. In addition, we design a mixed feature pyramid within the DST, which provides interlayer interaction information and intralayer multiscale information for a more complete feature learning within the new network. Furthermore, we impose a self-supervised strategy to guide the VGG16 provide the semantic change information from the output features of the encoder. We compared our experimental results with those of the state-of-the-art methods on four commonly used public VHR remote sensing datasets. It shows that our network performs better, in terms of F1, IoU, and OA, than those of the existing networks for change detection.

In the evolving domain of built-up area surveillance, remote sensing technology emerges as an essential instrument for Change Detection (CD). The introduction of deep learning has notably augmented the precision and efficiency of CD. This study focuses on the integration of deep learning methodologies, specifically the diffusion model, into remote sensing CD tasks for built-up urban areas. The goal is to explore the potential of a pre-trained Text-to-Image Stable Diffusion model for CD tasks and propose a new model called the Difference Guided Diffusion Model (DGDM). DGDM incorporates multiple pre-training techniques for image feature extraction and introduces the Difference Attention Module (DAM) and an Image-to-Text (ITT) adapter to improve the correlation between image features and text semantics. Additionally, DGDM utilizes attention generated from pre-trained Denoise UNet to enhance CD predictions. The effectiveness of the proposed method is evaluated through comparative assessments on four datasets, demonstrating its superiority over previous deep learning methods and its ability to produce more precise and detailed CD results. This innovative approach offers a promising direction for future research in urban remote sensing, emphasizing the potential of diffusion models in enhancing urban CD precision and automation. Our implementation code is available at https://github.com/morty20200301/cd-diffusion.

In recent years, numerous change detection methodologies have been proposed, with a predominant focus on binary change detection. Furthermore, there exists a paucity of research addressing semantic change detection in scenarios where solely binary change labels are available. This article introduces a multitask network for semantic change detection. First, 3-D ResUnet model is employed to generate initial multitemporal land cover results through postclassification comparison. Subsequently, the multitask network, encompassing two subtasks—binary change detection and multitemporal semantic segmentation—is proposed. Specifically, the shared branch of the network employs 3-D residual blocks to extract joint spectral-spatial features. In the subsequent task-specific branch, a 3-D GAN is incorporated for the binary change detection task to enhance the discrimination ability of latent high-level features for changes. Novel adaptive self-paced learning and certainty-weighted focal loss are proposed for multitemporal semantic segmentation to mitigate adverse effects from noisy semantic labels by considering sample complexity and reliability in the network optimization process. Experiments conducted on the Orbita Hyperspectral dataset in the Xiong'an New Area demonstrate the superior performance of the proposed method, achieving 99.28% and 76.60% for overall accuracy and kappa, respectively. This outperformance is notable when compared to other methods, such as Str4 and Bi-SRNet, showing an increase of 39.82% and 54.17% for kappa. Moreover, comparative experiments on SECOND further confirm the advantage of the proposed method, achieving 54.62% for kappa and outperforming other comparative methods, such as Bi-SRNet (47.61%).

Deep learning has shown superiority in change detection (CD) tasks, notably the Transformer architecture with its self-attention mechanism, capturing long-range dependencies and outperforming traditional models. This capability provides the Transformer with significant advantages in capturing global-level features of complex changes in objects within high-resolution remote sensing images. Though Transformers are mature in Natural Language Processing (NLP), their application in computer vision, particularly CD tasks, is nascent. Current research on leveraging Transformers for CD reveals limitations, especially under varied lighting and seasonal changes. To address this, we propose VisionTwinNet, a two-stage strategy. First, our Gated EnhanceClearNet, a specially designed deep network reduces image noise and enhances brightness, preserving shadows and correcting color distortions. With its unique gating mechanism, this network can adaptively adjust the importance of features, thereby exhibiting superior performance in various remote sensing image degradation issues. Secondly, we have developed Hybrid Light-Robust CDNet, a hybrid robust lightweight network custom-designed for CD in remote sensing images. This module deeply integrates the advantages of CNN and Transformer and introduces an innovative attention mechanism design, optimizing the key/value dimensions separately, instead of adopting traditional single linear transformations, ensuring efficient detection. Specifically, the LR-Transformer Block employs a lightweight multi-head self-attention mechanism, optimizing computational efficiency while providing richer feature representations. Comparative studies with six CD methods on three public datasets validate VisionTwinNet’s robustness and efficacy. Our approach notably reduces algorithmic complexity and enhances the efficiency of the model.

Deep-learning techniques have made significant advances in remote sensing change detection task. However, it remains a great challenge to detect the details of changed areas from high-resolution remote sensing images. In this study, we propose a full-scale difference feature fusion network (FDFF-Net) for change detection, which can alleviate pseudochanges and reduce the loss of change details during detection. In the encoding stage, a dense difference fusion module is proposed to effectively mine and fuse the multiple differences for each feature level between bitemporal images, leading to a substantial reduction in missed detection of change areas. Additionally, the different levels of difference features are aggregated through a full-scale skip connection, allowing the network to detect multiple changed objects with various sizes. In the decoding stage, a strip spatial attention module is designed to enhance the perception of the change areas, which improves the ability to detect detailed changes. The experiments on three change detection datasets, CDD, LEVIR-CD, and S2Looking, demonstrate that FDFF-Net outperforms the compared state-of-the-art methods and can detect more complete changes of small objects and clear contours of changed areas.

With increasingly rapid development of convolutional neural networks, the field of remote sensing has experienced a significant revitalization. However, understanding and detecting surface changes, which necessitate the identification of high-resolution remote sensing images, remain substantial challenges in achieving precise change detection. Excited deep learning-based change detection techniques often exhibit limitations and lack the necessary precision to detect edge details or other nuanced information in remote sensing images. To address these limitations, we propose a unique semantic segmentation deep learning network, the self-adaptive Siamese network (SASiamNet), specifically devised for enhancing change detection in remote sensing images. The SASiamNet excels in real-time land cover segmentation, adeptly extracting local and global information from images via the backbone residual network. Furthermore, it incorporates a primary feature fusion module to extract and fuse the primary stage feature map, and a high-level information refinement module to refine the resultant feature map. This methodology effectively transmutes low-level semantic information into high-level semantic information, thereby improving the overall detection process. Aimed at empirically testing the effectiveness of the SASiamNet, we utilize two distinct datasets: the public dataset, LEVIR-CD, and a challenging dataset, CDD. The latter is composed of bitemporal images sourced from Google Earth, spanning various regions across China. The experiment results unequivocally demonstrate that our approach outperforms traditional methodologies as well as contemporary state-of-the-art change detection techniques, hence underscoring the efficacy of the SASiamNet in the context of remote sensing image change detection.

As dust source bare soil land (BSL) contributes to air pollution and affects the photosynthesis of green plants and carbon absorption, it is the objective of this study to develop an approach for monitoring the changes of BSL using remote sensing technology. Unlike other land use/cover types, the classification of BSL as well as its change detection is often ignored. For traditional convolutional neural networks, deep layers cause a long range between input and output, inevitably leading to the loss of information and computational costs. To alleviate this problem, transformer is available to model the global dependencies. Bitemporal association, which is described as subtraction or attention mechanism, is not fully considered by current methods. Therefore, we proposed a spatial-temporal association enhanced mobile-friendly vision transformer (STAE-MobileVIT) for change detection of high-resolution images with light weight and high efficiency. On the one hand, a temporal association enhanced MobileVIT block is employed to strengthen the association of bitemporal images during feature extraction. On the other hand, a multiscale feature difference aggregator enhanced by spatial association is designed to fuse semantic and detailed information. Since the lack of binary change detection dataset for BSL, we established a small dataset named BSL-CD, consisting of 1083 pairs of 0.8 m bitemporal images with the size of 256 × 256 pixels, along with the corresponding labels. The experiments on BSL-CD show that our light-weight model surpass seven common methods by 3.48, 5.05, and 1.44 percent on F1, IoU, and OA, which proves the efficiency and accuracy of STAE-MobileVIT.

In the field of remote sensing, change detection is a crucial study area. Deep learning has made significant strides in the study of remote sensing image change detection during the past few years. Deep learning techniques still have some drawbacks. The global context cannot be modeled by convolutional neural networks due to the receptive field's restrictions. When extracting visual characteristics, the neural network does not concentrate more on the change region, which results in poor distinction between change and no-change regions. To address these problems, we propose networks with large receptive fields (LRFs) and difference image enhancement. First, we design the LRF strategy. It employs a long kernel shape in one spatial dimension for obtaining a long range of relations. Keeping a narrow kernel size in the other spatial dimension can extract local context information while avoiding interference from irrelevant regions. To focus on the changing features, we design the image difference enhancement (IDE) method, which decreases the distance between invariant features and enlarges the distance between changing features. In addition, we design the cross-channel interaction (CNI) strategy, which models the relationship between feature map channels and extracts feature representations through local CNI. On the CDD, WHU-CD, and LEVIR-CD public datasets, we conducted comprehensive experiments. According to the experimental results, our proposed LRDE-Net performs better than other state-of-the-art change detection techniques, and the change regions are more precisely identified. It can better cope with seasonal changes, light intensity, and other pseudochange disturbances.

Hyperspectral change detection (HCD) is widely applied in various domains, such as accurate agriculture, disaster assessment, land use, and environmental monitoring. Most HCD methods aim at extracting and classifying the spectral variation features with dimension reduction and machine-learning methods. Different from previous work, this letter proposes an edge-guided HCD method. Specifically, a subtraction operation is adopted to extract difference hyperspectral image (HSI). Then, the edge-preserving filtering is performed on the difference HSI to extract spectral–spatial features. Next, the number of the extracted features is diminished through the kernel principal component analysis (PCA). Finally, the fused features are input into a spectral classifier followed by the edge-preserving filtering to obtain the final change detection result. Experiments on several HCD datasets demonstrate that the proposed method can consistently outperform other advanced approaches in both subjective and objective evaluations when only a limited number of labeled samples are available. [ABSTRACT FROM AUTHOR]

Change detection of high-resolution remote sensing images is more attractive, since it can not only identify areas of changes but also identify types of changes. In this context, simultaneous change detection and semantic segmentation are natural and necessary. However, traditional methods put less emphasis on the cooperation of the above two tasks. In this letter, a novel method is proposed to realize the collaborative learning of change detection and semantic segmentation. By elaborately exploring the relevance and consistency between change detection and semantic segmentation, the proposed method synchronously enhanced feature separability of two tasks, and it outperformed a single change detection network or semantic segmentation network. Specifically, the proposed approach extracts multilevel bitemporal features by a backbone network, followed by two layer-by-layer decoders for learning change features and semantic features. On one hand, the interactive fusion module (IFM) fuses the changing features and semantic features together to increase the collaboration between the two tasks. On the other hand, the contrastive loss (CL) enhances the constraints between the two tasks. The advantages of the proposed method are demonstrated with respect to change region detection and change-type identification. [ABSTRACT FROM AUTHOR]

Synthetic aperture radar (SAR) image change detection is a critical task and has received increasing attention in the remote sensing community. However, existing SAR change detection methods are mainly based on convolutional neural networks (CNNs), with limited consideration of global attention mechanism. In this letter, we explore Transformer-like architecture for SAR change detection to incorporate global attention. To this end, we propose a convolution and attention mixer (CAMixer). First, to compensate for the inductive bias for Transformer, we combine self-attention with shift convolution in a parallel way. The parallel design effectively captures the global semantic information via self-attention and performs local feature extraction through shift convolution simultaneously. Second, we adopt a gating mechanism in the feed-forward network to enhance the nonlinear feature transformation. The gating mechanism is formulated as the element-wise multiplication of two parallel linear layers. Important features can be highlighted, leading to high-quality representations against speckle noise. Extensive experiments conducted on three SAR datasets verify the superior performance of the proposed CAMixer. The source codes will be publicly available at https://github.com/summitgao/CAMixer. [ABSTRACT FROM AUTHOR]

Remote sensing image change detection refers to finding the changed regions from a pair of registered images. It has important applications in many fields. However, most methods based on convolutional neural networks and transformers have high complexity and cannot be effectively deployed on satellites or drones in practical applications. To address this issue, we propose an extremely lightweight change detection algorithm called ELW_CDNet. Its inference speed is very fast. This method is based on the extremely lightweight shufflenetv2. Moreover, considering that both global as well as local features play an important role in change detection, we design a light global-local feature enhancement module (LGLFEM) for reinforcing the features extracted by the backbone. Specifically, the global feature extraction module in LGLFEM is implemented using separable self-attention (SSA), which has linear complexity and very low computational effort. We conduct experiments on two change detection datasets. Compared with some state-of-the-art methods, the proposed method can achieve superior performance with extremely fast inference speed. On the LEVIR-CD dataset, it achieves an F1 score of 90.47%, an intersection over union (IoU) of 82.60%, and a frame per second (FPS) of 914 with 1.75 M parameters and 1.91 GFLOPs. The code will be released soon on the site of https://github.com/dyl96/ELW_CDNet. [ABSTRACT FROM AUTHOR]

Deep learning techniques have shown increasing potential in change detection (CD) in hyperspectral images (HSIs). However, most deep learning-based existing methods for HSI CD follow a patch-based local learning framework and concentrate on binary CD. In this letter, we propose an end-to-end semantic feature fused global learning (SFFGL) framework for HSI multiclass CD (MCD). In SFFGL, a global spatialwise fully convolutional network (FCN), which introduces a spatial attention mechanism (PAM) between encoder and decoder, is designed to effectively exploit the global spatial information from the whole HSIs and achieve patch-free inference. PAM can adaptively extract global spatialwise feature representation. In the model training stage, a global hierarchical (GH) sampling strategy is introduced to obtain diverse gradients during backpropagation for more robust performance. The semantic–spatial feature fusion ($\text{S}^{2}\text{F}^{2}$) unit is designed to effectively fuse the enhanced spatial context information in the encoder and the semantic information in the decoder. More importantly, a semantic feature enhancement module (SEM) is proposed to weaken the influence of the unchanged regional background on the change regional foreground, thus further improving the accuracy. Experimental results on two benchmark HSI datasets demonstrate the effectiveness of the proposed SFFGL. [ABSTRACT FROM AUTHOR]

Remote sensing change detection aims to identify significant variations in aerial image acquisition during different time frames. A decisive change detector is necessary to filter out the interest regions, such as recent urban buildings and changed vegetation, from undesired detections, i.e., artifacts generated by misregistration and illumination changes. To overcome common change detection problems (false positive and false negative alarms) and also processing overhead, this manuscript proposes a lighTweIght temporal attention network, aka TITAN, which comprises a partial-siamese deep learning-based change detector that leverages the natural capacity of an encoder–decoder framework to extract different levels of feature information from its input data. To assist the process of combining the meaningful encoded spatial–temporal information with its corresponding semantic decoded counterpart, we also propose the temporal change attention module (TCAM). Although TCAM does not explicitly account for nonlocal spatial changes, results support the claim that it implicitly helped TITAN in the matter. The experimental results show the proposed approach overcomes three out of four state-of-the-art techniques in terms of overall average F-measure, Intersection over Union (IOU), and Percentage of Wrong Classification (PWC) measures calculated over SZATAKI, Onera, LEVIR, and SYSU-CD remote sensing change detection datasets, with the lowest overhead. [ABSTRACT FROM AUTHOR]

Lightweight change detection models are essential for industrial applications and edge devices. Reducing the model size while maintaining high accuracy is a key challenge in developing lightweight change detection models. However, many existing methods oversimplify the model architecture, leading to a loss of information and reduced performance. Therefore, developing a lightweight model that can effectively preserve the input information is a challenging problem. To address this challenge, we propose LightCDNet, a novel lightweight change detection model that effectively preserves the input information. LightCDNet consists of an early fusion backbone network and a pyramid decoder for end-to-end change detection. The core component of LightCDNet is the Deep Supervised Fusion Module (DSFM), which guides the early fusion of primary features to improve performance. We evaluated LightCDNet on the LEVIR-CD dataset and found that it achieved comparable or better performance than state-of-the-art models while being 10–117 times smaller in size. [ABSTRACT FROM AUTHOR]

Deep learning (DL) has recently achieved outstanding performance in change detection of multitemporal images. However, most existing DL-based change detection methods still suffer from the problem of insufficient labeled training samples. To overcome this limitation, an unsupervised superpixel-guided self-supervised learning network (S3Net) is proposed for detecting changes occurred on the land surface. By performing principal component analysis on two input images, a triple-channel pseudocolor image containing the main information of both the images is first generated, which is used for superpixel segmentation to produce homogeneous image objects. Then, a Siamese network composing of two identical subnetworks with shared weight based on transfer learning is trained for pretext task in a self-supervised learning way, aiming to obtain multiscale object-level spatial feature difference images. On this basis, a high-quality difference image is generated by incorporating the pixel-level and object-level difference information using a simple weighted fusion strategy, which can be analyzed by thresholding to produce the final binary change map. The experimental results on four real-world datasets from different sensors show that the proposed approach can obtain superior performance in comparison to several state-of-the-art change detection methods, which further demonstrates its effectiveness and practicability. We make our data and code publicly available (https://github.com/OMEGA-RS/S3Net_CD). [ABSTRACT FROM AUTHOR]

Synthetic aperture radar (SAR) image change detection is a hot but challenging task due to SAR images’ complex contents and inherent speckle noises. The expected change detection methods should reduce the influence of speckle noises, obtain the discriminative feature representations, and generate accurate change maps simultaneously. To these ends, we propose a new SAR image change detection method named feature fusion of information transfer network (FFITN). First, we develop a hybrid convolution block (HCB) to depress the speckle noise impacts and explore the valuable information from SAR images. Thus, the feature extraction module (FEM) is constructed to obtain the multilevel features. Then, an information transfer module (ITM) is proposed to capture the salient regions from various aspects. Also, the salient knowledge is transferred among features at different levels to enhance their discrimination. Next, a self-attention-based feature fusion module (SAFFM) is introduced to fuse various features. Finally, a change map generation module (CMGM) with the clustering algorithm and specific loss functions is designed to produce the pseudo labels and change maps. Experimental results on three public SAR datasets demonstrate the model’s effectiveness. Our source codes are available at https://github.com/TangXu-Group/FFITN. [ABSTRACT FROM AUTHOR]

Learning performance is unsatisfactory when training deep-learning networks without prior-knowledge guidance. In this letter, a multiscale change detection neural network guided by a change gradient image (CGI) was proposed. First, a multiscale information attentional module was embedded in the backbone of UNet to achieve a multiscale information fusion task of bitemporal images. Second, the position channel attention module (PCAM) was promoted to make the neural network pay more attention to the spectral and spatial information in the multiscale fused feature map. Finally, a change gradient guide module (CGGM) was proposed to optimize backpropagation and overcome the negative effects of pseudo-change. Compared with seven state-of-the-art methods using three pairs of real remote sensing images, the proposed approach could smoothen the salt-and-pepper noise from the detection maps and improve the detection accuracy. The quantitative improvements are about 1.67% and 3.00% in terms of overall accuracy (OA) and kappa coefficient, respectively, thus confirming the feasibility and superiority of the proposed approach for detecting land cover change with remotely sensed images. Code: https://github.com/ImgSciGroup/MACGGNet.git. [ABSTRACT FROM AUTHOR]

Change detection based on synthetic aperture radar (SAR) images is an important application in the remote-sensing technology field. However, the lack of labeled data has been a difficult problem in SAR image detection, especially for pixel-level change detection. In this letter, we propose a novel unsupervised change detection algorithm, which improves the detection accuracy by exploring features from both spatial and frequency domains of SAR images. In particular, first clustering is used as preclassification to obtain pseudo-labels and then by incorporating classifiers and pseudo-labels in terms of feature learning, a novel unsupervised detection algorithm is proposed. To improve the sensitivity of the algorithm to changed details and enhance the antinoise ability of the change detection network, the attention mechanism (AM) is integrated into the network to fully extract important spatial structure information. Moreover, a multidomain fusion module is proposed to integrate spatial and frequency domain features into complementary feature representations. This module contains multiregion features weighted by the channel-spatial AM and deep features filtered out by the gated linear units (GLUs) in the frequency domain. To verify the effectiveness of the proposed algorithm, it is compared against the other four SAR image change detection algorithms using three real datasets. The experimental results show that the proposed method outperforms the other four algorithms in terms of percent correct classification (PCC) and Kappa coefficient (KC). [ABSTRACT FROM AUTHOR]

High-resolution remote-sensing image change detection plays an important role in areas such as land and resources investigation, natural disaster prediction, and military strategy research. Current change detection methods often focus on extracting more discriminative features, while ignoring the information loss and imbalance problems in the process of feature fusion, which results in weakness in small change objects and edge pixels of change objects. In this letter, a simple but efficient network architecture, extraction, comparison and fusion network (ECFNet), for change detection in remote-sensing images is proposed. By constraining the number of feature channels in the fusion process by the feature comparison module (FCM), ECFNet can better utilize the fine-grained information in the multiscale feature map for result prediction, which not only improves the detection performance of small objects, but also reduces the false detection around the edge pixels of change objects. Experiments on the deeply supervised image fusion network for change detection (DSIFN-CD) test set show that ECFNet achieves state-of-the-art results with a small amount of computation. [ABSTRACT FROM AUTHOR]

Following the intuitive idea of detecting changes by directly measuring dissimilarities between pairs of features, change detection methods based on feature similarity learning have emerged as a crucial field. However, large variances in the scale and location of required contextual information and heavy imbalance between easy and hard samples remain challenging issues. To address the first issue, we propose the Local-Specificity and Wide-View Attention Network (LSWVANet), which features a series of attention modules named Local-Specificity and Wide-View Attention Modules (LSWVAMs). Each LSWVAM consists of two contextual feature units: the Local-Specificity Feature Pyramid unit, which extracts part-specific contexts at the fine-grained level to focus on subtle changes within local discriminative parts, and the Wide-View Feature Pyramid unit, which extracts wide-view contexts at the long-range level to highlight significant changes in large-scale regions. To tackle the second issue, we introduce a novel sample-specific loss function called Hard Sample-Aware Contrastive Loss (HSACL), which is designed to downweight easy samples from both changed and unchanged categories, thereby rapidly shifting the training focus towards the informative hard samples. We demonstrate the effectiveness of our method through experiments on three challenging datasets, VL-CMU-CD, PCD2015 and PSCD, and report the experimental results showing that our approach achieves state-of-the-art accuracy.

The world's forests are experiencing rapid changes due to land use and climate change. However, a detailed map of global forest gain at fine spatial and temporal resolutions is still missing. To fill this gap, we developed an automatic framework for mapping annual forest gain globally using Landsat time series, the Landsat-based detection of trends in disturbance and recovery (LandTrendr) algorithm, and the Google Earth engine platform. First, stable forest samples collected based on the first all-season sample set and an automated sample migrate method were used to determine annual normalized burn ratio (NBR) thresholds for forest gain detection. Second, with the NBR time series from 1982 to 2021 and LandTrendr algorithm, we produced a dataset of global forest gain year from 1983 to 2021 based on a set of decision rules. Our results reveal that over 60% gains occurred in Russia, Canada, the United States, Indonesia, and China, and approximately half of global forest gain occurred between 2001 and 2010. The forest gain map developed in this study exhibited good consistency with statistical inventories and independent regional and global products. Our dataset can be useful for policy-relevant research on the global carbon cycle, and our method provides an efficient and transferable approach for monitoring other types of land cover dynamics.

Change detection refers to revealing the surface changes from multitemporal images of a given scene, where changed regions of different size and shape may appear anywhere. Convolutional neural network, as one of the most widely used deep learning architectures, has shown good performance in change detection task recently. However, due to multiple convolution and pooling operators in deep architectures, it often happens that small changes are missed and sharp edges are blurred. In this article, an adaptive spatial and difference learning method is proposed to tackle this problem, which mainly contains a feature-adapted difference learning (FADL) module and a difference recalibration (DR) module. In FADL, a kernel-adapted spatial learning part is constructed to capture varying spatial information, which selects proper kernels to adapt different shape and size of changed regions; a joint feature refinement part is employed to learn comprehensive features between bitemporal data, which are further enhanced by difference information. In addition, DR with shallow features is designed to recalibrate spatial and difference details further. From the experiments on three public datasets, our proposed method can relieve the spatial ambiguity problem and obtains the optimal results among the compared change detection methods.

Remote sensing image change detection (CD) refers to the automated or semi-automated detection of differences between two remote sensing images taken at different times in the same region. To achieve better global modeling and faster inference, we propose a network architecture containing a hierarchical swin transformer block and deformable attention transformers crossed for encoding and lightweight MLP decoding to solve the CD task. The deformable attention transformer allows adaptive adjustment of the relationships and weights between feature mappings to effectively combat variations and noise interference in various scenes. The alternating use of swin transformer block and deformable attention transformer ensures the efficiency as well as the flexibility of the model. The lightweight MLP approach provides better ability to extract spatial features and contextual information, as well as faster inference speed. Compared with other methods, our proposed DAHT-Net method improves F1 scores by 0.98 and 2.61 on LEVIR-CD, CDD and two publicly available benchmark datasets, respectively, and performs well on other measures. These experimental results validate that the DAHT-Net network outperforms other comparative methods and highlight its effectiveness in remote sensing image change detection. In summary, our proposed hierarchical deformable attention-guided transformer network model provides a promising solution for remote sensing image change detection with superior performance compared to other state-of-the-art methods.

In recent years, the important evolution in the number, potentiality, and diversity of Earth observation (EO) satellites has resulted in dramatic increases in the payload data volume and rate. However, these exponential increases in the generated data volume are creating a significant bottleneck onboard EO satellites due to transmission bandwidth limits and communication delays. Onboard imaging payload data processing can provide an appropriate solution to alleviate the induced data bottleneck. It can also facilitate rapid response for decision-making operations. Change detection is one of the most significant functions in onboard payload data processing systems that enable a real-time reaction to natural disasters, such as flooding, earthquakes, and volcanic eruptions. In this article, we address the problem of automatic change detection onboard EO satellites. This article aims to design an automatic onboard change detection system (OCDS) that can run on existing flight-proven hardware by taking advantage of the attractive features of a leading model in deep learning (DL) called convolutional neural network. The contribution of this article is twofold. An efficient algorithmic solution for change detection based on DL that fulfills space environment-induced constraints is first proposed. Second, a preliminary hardware architecture of the proposed OCDS is designed based on payload data processing flight-proven hardware. The experimental results demonstrate the efficiency of the proposed DL-based change detection approach and the suitability of the designed OCDS for onboarding on EO small satellites.

Synthetic aperture radar (SAR) image change detection (CD) focuses on identifying changes between two images at different times for the same geographical region. Recently, several deep learning methods have been proposed for performing SAR based CD. However, speckle noise remains a major challenge for these methods. To address this, we propose a despeckling model (DM) that effectively suppresses speckle noise and enhances the performance of the existing CD methods. The proposed despeckling architecture is not only resilient to multi-temporal SAR acquired from one SAR imaging process (i.e., the same number of SAR images looks before and after the change) but also deals with any combination of single or multi-look images acquired prior and after the change. Moreover, as a second contribution, we have also proposed a loss function that effectively suppresses speckle noise, thereby improving the change detection accuracy. Both the despeckling model and the proposed tolerant noise loss function are evaluated extensively on three public real SAR datasets, achieving superior performance compared to existing state-of-the-art SAR CD methods in all datasets.

Radiometric consistency of multitemporal satellite observations is affected by sensor stability and scene related issues. Relative radiometric normalization (RRN) is a widely used method to reduce these radiometric differences, its performance depends on the accurate identification of representative pseudoinvariant features (PIFs). However, existing RRN methods are mainly developed for bitemporal images and are limited to time-series imageries due to the complexity of identifying effective PIFs. In this study, we proposed a novel RRN method to enhance the radiometric consistency of Landsat time-series imageries. This method includes the following: first, a trend-based PIFs identification considering land cover changes and phenological trends from the entire time series; second, a PIFs optimization involving an automatic reference selection and a PIFs refining for each reference–target image pair; and third, a combined RRN modeling using the M-estimator sample consensus algorithm and robust linear regression. The Landsat surface reflectance products were used to validate the proposed method. The experimental results showed that the trend-based PIFs identification provided the consistent PIFs for all reference–target image pairs; aided by an automatic reference allocation, PIFs optimization filtered the proper PIFs with high spectral and spatial similarity for each image pair in monthly image stack; the proposed RRN method achieved good performance in model precision and radiance consistency improvement; the proposed RRN method outperformed seven commonly used RRN methods on majority images in image stack of December. The normalized images can help generate more comparable time-series analysis results by reducing the uncertainties from radiometric calibration, atmospheric correction, and sensor differences.

One of the major challenges in the change detection (CD) of high-resolution remote sensing images is the high requirement for computational resources. Besides, to get the best change detection result, it must spot only the important changes while omitting unimportant ones, which requires learning complex interactions between multi-scale objects on the images. Despite Convolution Neural Network (CNN) efficiently extracting features from such images, it has a limited receptive field resulting in sub-optimal representation. On the other hand, Vision Transformer (ViT) can capture long-range dependencies. Still, it suffers from quadratic complexity concerning the number of image patches, especially for high-resolution images. Furthermore, both approach can not model the interactions among multi-scale image patches, which is essential for a model to fully understand the natural images. We propose FocalCD, a CD method based on a recently proposed focal modulation architecture capable of learning short and long dependencies to solve this problem. It is attention-free and does not suffer from quadratic complexity. Also, it supports learning multi-scale interaction by adaptively selecting the discriminator regions from multi-scale levels. Besides the efficient yet powerful encoder, FocalCD has an effective multi-scale feature fusion and pyramidal decoder network. FocalCD achieves strong empirical results on various CD datasets, including CDD, LEVIR-CD, and WHU-CD. It reaches F1 scores of 0.9851, 0.952, and 0.9616 on datasets CDD, LEVIR-CD, and WHU-CD outperforming state-of-the-art CD methods while having comparable or even lower computation complexity.

Satellite image time-series change detection methods have become an effective means of obtaining information on land cover change. However, the temporal, spectral, and spatial features and their derived features of land cover objects are of great importance for time-series change detection. The existing studies have made insufficient use of these features, which may affect the results of land cover change detection. In order to fully integrate the above features to portray and represent change information, this study proposes a coupled temporal–spectral–spatial multidimensional information change detection framework (TSSF) method. First, the derived index features are calculated to construct intra-annual temporal–spectral information to reduce the underutilization of spectral features. Second, the intra-annual temporal–spectral information is extended to the spatiotemporal domain by the simple noniterative clustering method and the SG filtering method to increase the exploitation of spatial features. Then, the value and shape-based dynamic time warping method and the change vector analysis in the posterior probability space method are employed to obtain change information from the spectral, index, and class probability perspectives. Finally, the change type in the change region is obtained from the class probability information of the change magnitude according to the Bayesian criterion. Tianjin City was used as the study area to explore the land cover change from 1990 to 2020. The results show that the TSSF method is feasible in expressing temporal–spectral–spatial change information compared with the existing methods and is conducive to the efficient acquisition and identification of change areas and change types.