Multiaperture echo signals collected by distributed array SAR can be reconstructed to realize High-Resolution and Wide-Swath (HRWS) imaging. To achieve the HRWS imaging of distributed SAR, an advanced Azimuth Multi-Aperture Reconstruction (AMAR) method is put forward for the first time. First, to eliminate the trajectory offset of distributed SAR, the first-order baseline deviation, second-order baseline deviation and azimuth space variation error are compensated in turn for the calibrated and undersampled echo signals, and the echoes containing distorted azimuth ambiguity can be obtained. Second, the distortion of ambiguity is corrected by range decompression, anti-migration correction, etc. Finally, the echoes satisfying the Nyquist theorem are derived by the AMAR of Train formation, and the bistatic SAR imaging is performed. In addition, some simulations are conducted to evaluate the performance of the proposed method, and the results demonstrate that the method can achieve AMAR and HRWS imaging of distributed array SAR.

Unsupervised change detection using contrastive learning has significantly improved the performance of literature techniques. However, at present it only focuses on the bi-temporal change detection scenario. Previous state-of-the-art models for image time-series change detection have traditionally depended on features obtained either through clustering learning or by training models from scratch using pseudo labels tailored to each scene. However, these approaches fail to either exploit the spatial-temporal information of image time-series or generalize to unseen scenarios. In this work, we propose a two-stage approach to unsupervised change detection in satellite image time-series using contrastive learning with feature tracking. By deriving pseudo labels from pre-trained models and using feature tracking to propagate them within the image time-series, we improve the consistency of our pseudo labels and address the challenges of seasonal changes in long-term remote sensing image time-series. We adopt the self-training algorithm with ConvLSTM on the obtained pseudo labels, where we first use supervised contrastive loss and contrastive random walks to further improve the feature correspondence in space-time. Then a fully connected layer is fine-tuned on the pre-trained multi-temporal features for generating the final change maps. Through comprehensive experiments on two datasets, we demonstrate consistent improvements in accuracy on fitting and inference scenarios.

The mechanism of connecting multimodal signals through self-attention operation is a key factor in the success of multimodal Transformer networks in remote sensing data fusion tasks. However, traditional approaches assume access to all modalities during both training and inference, which can lead to severe degradation when dealing with modal-incomplete inputs in downstream applications. To address this limitation, we propose a novel approach to incomplete multimodal learning in the context of remote sensing data fusion and the multimodal Transformer. This approach can be used in both supervised and self-supervised pre-training paradigms. It leverages the additional learned fusion tokens in combination with modality attention and masked self-attention mechanisms to collect multimodal signals in a multimodal Transformer. The proposed approach employs reconstruction and contrastive loss to facilitate fusion in pre-training, while allowing for random modality combinations as inputs in network training. Experimental results show that the proposed method delivers state-of-the-art performance on two multimodal datasets for tasks such as building instance / semantic segmentation and land-cover mapping when dealing with incomplete inputs during inference.

Effective mitigation of atmospheric phase screen (APS) is crucial for synthetic aperture radar (SAR) interferometry in coastal regions. However, existing methods suffer from various limitations in estimating atmospheric delays, especially turbulent mixing components. As a solution, we propose a sparse-convolution-based attention deep residual U-shaped network (sARU-Net), a submanifold sparse convolutional neural network (SSCNN), for adaptive estimation of APS from InSAR coherent pixels. Moreover, we propose a straightforward and effective iterative processing strategy for generating sample datasets for model training. The results obtained from the TerraSAR-X and Sentinel-1 datasets in two coastal regions of China demonstrate a reduction in the standard deviation of the interferograms by 77.7% and 68.3%, respectively, after applying our correction method. In addition, the InSAR results align more closely with the leveling data. On the same dataset, our method shows superior performance compared with both the generic atmospheric correction online service (GACOS) for InSAR method and the linear model method. When compared with previous convolutional neural network (CNN), (i.e., attention-based deep residual U-shaped network, ARU-Net) method, our approach exhibits better accuracy and more detail and is 21% more computationally efficient for training, while requiring 24% less graphics processing unit (GPU) memory. The effectiveness of the proposed method shows that sparse convolution (SC) has great potential in mitigating the effects of APS in InSAR coherent pixels.

Change detection (CD) in high-resolution remote sensing has received large attention due to its wide range of applications. Many methods have been proposed in the literature and achieved excellent performance. However, they are often fully supervised, thus requiring abundant pixel-level labeled samples, which is time-consuming and labor-intensive. Especially compared to the common single-temporal interpretation, labeling bi-temporal images is often more complicated. Therefore, this study combines weakly supervised learning (WSL) to reduce label acquisition costs. But changed regions are small, fragmented, and similar to the background, which increase the gap between weakly supervised and fully supervised tasks. To address these difficulties, we explore self-supervised methods to construct a WSL framework based on image-level labels for general CD, termed WSLCD in this paper. First, we design a double-branch siamese network to derive embeddings and initial class attention maps (CAMs), which inputs the original image pair and the spatially transformed image pair. Second, mutual learning and equivariant regularization (MLER) is enforced on CAMs from different views, which implements consistency constraints in confusion regions and makes CAMs learn from each other based on saliency regions. Furthermore, prototype-based contrastive learning (PCL) is designed such that unreliable pixels can learn from prototypes computed from reliable pixels. PCL includes intra-view contrast and cross-view contrast depending on whether the prototypes and class embeddings are from the same view. With the above strategies, we narrow the gap between image-level weakly supervised CD and fully supervised CD. Experiments are conducted on three CD datasets, including CLCD, DSIFN and GCD. Our method achieves state-of-the-art performance on pseudo label generation and CD. The code is available at https://github.com/mfzhao1998/WSLCD.

In this work, we address the challenge of open-world semantic segmentation for remote sensing (RS) images, which involves segmenting arbitrary objects in images using open RS data. Previous efforts in open-world segmentation mostly focus on Internet-scale paired image-text data with rich vocabulary of concepts. However, these works cannot be directly transferred to RS domain due to the lack of large-scale RS data-text pairs and the corresponding annotations. To overcome this limitation, we propose using text descriptions and annotations from OpenStreetMap as a source of supervision while using images from satellite images. We utilize a conditional Unet model to predict segmentation masks given a text description, and leverage the rich information contained in a pretrained CLIP model to align the images and the corresponding text embeddings using a contrastive loss. Our experimental results demonstrate the potential of open-world segmentation on open RS data.

SAR and optical images provide complementary information on land-cover categories in terms of both spectral signatures and dielectric properties. This paper proposes a new unsupervised land-cover segmentation approach based on contrastive learning and vector quantization that jointly uses SAR and optical images. This approach exploits a pseudo-Siamese network to extract and discriminate features of different categories, where one branch is a ResUnet and the other branch is a gumble-softmax vector quantizer. The core idea is to minimize the contrastive loss between the learned features of the two branches. To segment images, for each pixel the output of gumble-softmax is discretized as a one-hot vector and its proxy label is chosen as the corresponding class. The proposed approach is validated on a subset of DFC2020 dataset including six different land-cover categories. Experimental results demonstrate improvements over the current state-of-the-art techniques and the effectiveness of unsupervised land-cover segmentation on SAR-optical image pairs.

There are limited studies on the semantic segmentation of high-resolution polarimetric synthetic aperture radar (PolSAR) images due to the scarcity of training data and the complexity of managing speckle noise. The Gaofen contest has provided open access a high-quality PolSAR semantic segmentation dataset. Taking this opportunity, we propose a multipath residual network (MP-ResNet) architecture for the semantic segmentation of high-resolution PolSAR images. Compared to conventional U-shape encoder–decoder convolutional neural network (CNN) architectures, the MP-ResNet learns semantic context with its parallel multiscale branches, which greatly enlarges its valid receptive fields and improves the embedding of local discriminative features. In addition, MP-ResNet adopts a multilevel feature fusion design in its decoder to effectively exploit the features learned from its different branches. Comparisons with the baseline method of fully connected network (FCN with ResNet34) show that the MP-ResNet has achieved significant accuracy improvements. It also surpasses several state-of-the-art methods in terms of overall accuracy (OA),$\text{m}F_{1}$and frequency weighted intersection over union (fwIoU), with only a limited increase of computational costs. This CNN architecture can be used as a baseline method for future studies on the semantic segmentation of PolSAR images. The code is available at:https://github.com/ggsDing/SARSeg.

The large amount of unlabeled remote sensing acquired from different sources and at different times (defined as multiple views in this article) presents both an opportunity and a challenge for change detection. Recently, many generative model-based methods have been proposed for remote sensing image change detection on such unlabeled data. However, the high diversities in the learned features weaken the discrimination of the relevant change indicators in unsupervised change detection tasks. Moreover, these methods lack research on massive archived images. In this work, a self-supervised change detection approach based on an unlabeled multiview setting is proposed to overcome this limitation. This is achieved by the use of a multiview contrastive loss in the feature alignment between multiview images. In this approach, a pseudo-Siamese network is trained to regress the output between its two branches pretrained in a contrastive way on a large dataset of single-sensor or cross-sensor image pairs. Finally, the feature distance between the outputs of the two branches is used to define a change measure, which can be analyzed by thresholding to get the final binary change map. Experiments are carried out on two single-sensor and three cross-sensor datasets. The proposed approach is compared with other supervised and unsupervised state-of-the-art change detection methods. Results demonstrate both improvements over state-of-the-art unsupervised methods and the proposed approach narrows the gap between unsupervised and supervised change detection.

The effective combination of the complementary information provided by huge amount of unlabeled multisensor data (e.g., synthetic aperture radar (SAR) and optical images) is a critical issue in remote sensing. Recently, contrastive learning methods have reached remarkable success in obtaining meaningful feature representations from multiview data. However, these methods only focus on image-level features, which may not satisfy the requirement for dense prediction tasks such as land-cover mapping. In this work, we propose a self-supervised framework for SAR-optical data fusion and land-cover mapping tasks. SAR and optical images are fused by using a multiview contrastive loss at image level and super-pixel level according to one of those possible strategies: in the early, intermediate, and late strategies. For the land-cover mapping task, we assign each pixel a land-cover class by the joint use of pretrained features and spectral information of the image itself. Experimental results show that the proposed approach not only achieves a comparable accuracy but also reduces the dimension of features with respect to the image-level contrastive learning method. Among three fusion strategies, the intermediate fusion strategy achieves the best performance. The combination of the pixel-level fusion approach and the self-training on spectral indices leads to further improvements in the land-cover mapping task with respect to the image-level fusion approach, especially with sparse pseudo labels. The code to reproduce our results will be found athttps://github.com/yusin2it/SARoptical_fusion.

Deep-learning techniques have achieved great success in remote-sensing image change detection. Most of them are supervised techniques, which usually require large amounts of training data and are limited to a particular application. Self-supervised methods solve these problems and are widely used in unsupervised binary change detection tasks. However, the existing self-supervised methods in change detection are suboptimal for pixel-wise change detection tasks. In this work, a pixel-wise contrastive approach is proposed to overcome this limitation. This is achieved by using contrastive loss in superpixel-level features on an unlabeled multiview setting. In this approach, a pseudo-Siamese network is trained to obtain pixel-wise representations and to align features from shifted image pairs. The final binary change map is obtained by using thresholding methods on learned temporal features. To overcome the season-related noise in binary change maps, we also used an uncertainty method to enhance the temporal robustness of the proposed approach. Two homogeneous (OSCD and MUDS) datasets and one heterogeneous (California Flood) dataset are used to evaluate the performance of the proposed approach. Results demonstrate improvements in both efficiency and accuracy over the patch-wise multiview contrastive method.

The availability of multi-sensor data presents an opportunity for change detection based on the complementary use of properties associated with different data sources. This paper proposes a new unsupervised change detection framework based on the joint use of SAR and optical images. The framework exploits a contrastive learning algorithm and the assumption of the scarcity of relevant changes. The proposed architecture is a pseudo-Siamese network, which is trained to regress the feature vector of bi-temporal concatenated SAR-optical input data. The output feature vectors from two branches of the pseudo-Siamese network are used to calculate change intensity maps. Then, the binary change map is obtained by setting a proper threshold. The proposed method is validated by using a multi-sensor dataset made up of Sentinel-1 and Sentinel-2 images that is also compared with the single use of each modality image. Experimental results demonstrate improvements of the multisensor approach over single modality and confirm the potentiality of the joint use of SAR and optical images in change detection.

Atmospheric phase screen (APS) is a very critical issue for the application of interferometric synthetic aperture radar (InSAR) techniques. The spatial-temporal variations of APS are the dominant error source in interferograms and may completely mask displacement signals. Many external meteorological data-based methods and phase-based methods have been developed in the past decades, but all have their inherent limitations. In this article, we propose a deep learning-based method, which is based on an attention-based deep residual U-shaped network (ARU-Net), to mitigate atmospheric artifacts. With this approach, APS patches and clean interferogram patches are sampled from InSAR interferograms to train the network. After training, the network can be used to mitigate the APS for individual interferograms. Compared with the generic atmospheric correction model (GACOS) and the advanced time-series InSAR method distributed scatterer interferometry (DSI), the key advantage of our method is that atmospheric delay can be effectively learned and removed from individual high-resolution interferometric phase itself without external data. Accuracy was validated by using individual and stacked interferograms from TerraSAR-X data over the Hong Kong International Airport (HKIA) and Hong Kong Science Park (HKSP) sites. The results showed that our method consistently delivered greater standard deviation (SD) reduction after APS correction than the GACOS method. Moreover, the time-series results were in agreement with the DSI and leveling measurements. The effectiveness of the proposed ARU-Net to remove APS effects from interferograms shows great potential for the development of a new set of deep learning-based APS reduction methods.