Compared with natural images, aerial images collected by satellite sensors/aerial cameras can provide a much larger field of view and often contain multiple objects of interest (multiple labels). There are certain limitations of existing multilabel aerial image classification methods. First, label correlations were often ignored in previous MAIC work, and thus, multilabel classifiers failed to be self-adapted. Second, existing multilabeled data sets for aerial images only cover limited images with fixed labels. Therefore, the underlying semantic correlations of labels cannot be fully included, while such correlation information is implicitly used as common knowledge by human beings. To tackle these concerns, we propose a novel multilabel classification method for aerial images. Our contributions are twofold. First, as the first attempt, label correlations are inferred from both the specific data set and ConceptNet (a popular knowledge graph for common sense). Second, based on graph neural network (GNN), we propose a novel end-to-end aerial image classification model, named the multiple label concept graph (ML-CG). ML-CG builds a concept graph to describe the semantic correlations from both the label set and the ConceptNet. We also incorporate both semantic attention and label attention in the GNN to better extract meaningful information of image labels. Compared with state-of-the-art methods, the effectiveness of the proposed method is demonstrated on both the commonly used UCM data set and a recently proposed DFC15 data set with high image resolution.

Deep learning (DL) methods are promising for the multilabel aerial image classification (MAIC) task. However, current DL methods face a common problem: the need for large multilabeled datasets. Collecting and annotating raw aerial image datasets can be extremely time- and labor-consuming. To address this concern in MAIC, domain adaptation (DA) provides a novel solution by transferring the knowledge learned from a label-rich dataset (i.e., the source domain) to a label-scarce dataset (i.e., the target domain), while current DA models are mainly designed for single-labeled tasks. In this article, we propose a novel end-to-end MAIC model based on DA techniques, named DA-MAIC. To the best of our knowledge, this article for the first time integrates DA to tackle the label scarcity problem in the MAIC task. Specifically, the proposed DA-MAIC is composed of two main parts: the image classifier and the domain classifier. The image classifier captures task-discriminative features based on the graph convolutional network (GCN) to predict multiple image labels; and the domain classifier extracts domain-invariant representations, which mitigates the domain shift between two underlying distributions. We extensively evaluate the proposed DA-MAIC from different perspectives on three benchmark datasets, including the commonly used UCM dataset, the high-resolution AID dataset, and the recently proposed DFC15 dataset. Both quantitative and qualitative results support that the proposed DA-MAIC can generalize the source domain knowledge to new scenarios and substantially improve the classification performance on the target domain task.

Aerial images play a key role in remote sensing as they can provide high-quality surface object information for continuous communication services. With advances in UAV-aided data collection technologies, the volume of aerial images has been greatly promoted. To this end, semantic understandings for these images can significantly improve the quality of service for smart devices. Recently, the multilabel aerial image classification (MAIC) task has been widely researched in academics and applied in industries. However, existing MAIC methods suffer from suboptimal performance as objects are located in different sizes and scales. To address these issues, we propose a novel multigrained semantic grouping model for aerial image learning, named MSGM. First, image features presented by the backbone are sent to spatial pyramid convolutional layers which extract the instances in a parallel manner. Then, three grouping mechanisms are designed to integrate the instances from the pyramid framework. In addition, MSGM builds a concept graph to represent the label relationship. MSGM resorts to the graph convolutional network to learn the concept graph directly. We extensively evaluate MSGM on two benchmark aerial image datasets, the commonly used UCM dataset, and the high-resolution DFC15 dataset. Quantitative and qualitative results support the effectiveness of the proposed MSGM.

We study Graph Neural Networks (GNNs)-based embedding techniques for knowledge graph (KG) reasoning. For the first time, we link the path redundancy issue in the state-of-the-art path encoding-based models to the transformation error in model training, which brings us new theoretical insights into KG reasoning, as well as high efficacy in practice. On the theoretical side, we analyze the entropy of transformation error in KG paths and point out query-specific redundant paths causing entropy increases. These findings guide us to maintain the shortest paths and remove redundant paths for minimized-entropy message passing. To achieve this goal, on the practical side, we propose an efficient Graph Percolation process motivated by the percolation phenomenon in Fluid Mechanics, and design a lightweight GNN-based KG reasoning framework called Graph Percolation Embeddings ( GraPE ). GraPE outperforms state-of-the-art methods in both transductive and inductive reasoning tasks, while requiring fewer training parameters and less inference time.

Oriented remote sensing object detection (ORSOD) remains challenging because objects usually appear at arbitrary rotations, with large aspect-ratio variations and dense spatial layouts. Although recent detectors have achieved notable progress, most still rely on standard convolutional backbones, whose limited inherent rotation equivariance weakens robustness to rotated targets and reduces localization reliability. To address these issues, this paper proposes RedhDet, an ORSOD framework based on rotation-equivariant feature learning and dual-head decoupling. Specifically, a Ladder-Structured Implicitly Equivariant Network (LSIEN) is developed as the backbone to extract rotation-equivariant features in a soft manner, without relying on strictly constrained kernels or equivariant pre-trained weights. In addition, a decoupled dual-head architecture is introduced, in which a fully connected branch is used for classification and a convolutional branch is used for bounding box regression, thereby better matching the distinct requirements of semantic recognition and precise localization. Experimental results on the DOTA-1.0 and HRSC2016 datasets demonstrate the effectiveness of the proposed method. On the DOTA-1.0 test set, RedhDet improves mAP₅₀ by 2.09% and mAP₇₅ by 14.84% over the baseline. On HRSC2016, gains of 14.12% and 51.85% are achieved for mAP₅₀ and mAP₇₅, respectively. With multi-scale training and testing, the proposed method achieves 80.40% mAP on DOTA-1.0, indicating its effectiveness and competitiveness for ORSOD.

Masked auto-encoders (MAEs) have been extensively employed in the field of semi-supervised hyperspectral image classification (HSIC). However, the developed models encounter significant challenges in learning separable representations, as they do not sufficiently prioritize the reconstruction of the central pixel, which hinders their ability to learn separable representations. To address this limitation, we propose spectral–spatial MAE with central pixel reconstruction (SSMAE-CR), a novel self-supervised framework tailored for HSIC. To capture more comprehensive representations, SSMAE-CR employs a dual-branch architecture comprising a spectral MAE with central pixel reconstruction (SpecMAE-CR) and a spatial MAE with central pixel reconstruction (SpatMAE-CR). SpecMAE-CR highlights the significance of central pixel reconstruction by measuring the deviation between the central pixels of reconstructed and original samples. To preserve the holism of the learned latent representations, SpatMAE-CR maps the central pixels of the reconstructed samples back to their original counterparts through the introduction of an additional linear layer. Rigorous comparative experiments conducted on four publicly available datasets fully demonstrate that SSMAE-CR outperforms state-of-the-art methods. Furthermore, we validate the effectiveness of SSMAE-CR by evaluating the mean intra-class and inter-class distances of the learned representations. Experimental results demonstrate that prioritizing central pixel reconstruction yields a statistically significant increase in the mean inter-class distance, suggesting enhanced class separability in the representation space.