Zhaoxiang Zhang

Reliable unmanned aerial vehicle (UAV) geolocation is essential for autonomous navigation, particularly in environments where satellite signals are vulnerable to spoofing and interference. However, current vision-based UAV positioning approaches rely heavily on visible imagery and lack large-scale multimodal datasets with strict spatial alignment, limiting their robustness across diverse scenes and temporal conditions. To bridge this gap, we introduce M3T-UAV, the first large-scale dataset specifically designed for vision-based UAV geolocation across multimodal, multiscene, and multitemporal scenarios. M3T-UAV contains more than 29,000 (29K) densely sampled and strictly registered triplets of visible, infrared (IR), and satellite images, covering seven representative natural scene attributes. Moreover, we extend the standard AUC@T metric and propose two novel evaluation measures, GeoAUC@T and Visual Positioning Score (VPS), to better assess multimodal UAV geolocation performance. Methodologically, we present LoFTDF, a semidense matching framework that achieves high-precision image matching by redesigning the soft assignment matrix learning procedure and introducing a dual-deformation field (DDF) module for fine-level geometric alignment. The proposed LoFTDF achieves state-of-the- art performance on the M3T-UAV benchmark, reaching 0.9036 VPS on the visible–satellite image matching task and 0.4421 VPS on the IR–satellite task, substantially outperforming existing matching-based geolocation methods. The code and dataset will be publicly available at: https:// github.com/JinHaiXu0322/M3T-UAV.

High-Resolution Range Profile (HRRP), which contains rich discriminative information about targets, is widely used in radar automatic target recognition. However, the aspect sensitivity of HRRP data limits recognition performance in scenarios with missing or varying aspects, while noise and other disturbances also distort HRRP waveforms, impairing recognition. To address these issues, inspired by causal theory, this paper proposes an HRRP target recognition framework comprising an Unsupervised Dynamic-Attribute Feature Disentanglement Network (UDAFD-Net) and an HRRP Recognition Network (HRRP-RecNet) in cascade. UDAFD-Net employs an asymmetric variational inference architecture to decompose HRRP sequence samples into dynamic features (related to aspect variations, noise, and other disturbances) and attribute features (solely linked to target types). A counterfactual regularization constraint is integrated into the Evidence Lower Bound (ELBO) loss to block spurious correlations between dynamic and attribute features, ensuring their independence. Finally, the disentangled attribute features and their corresponding type labels are used to train HRRP-RecNet, ensuring that recognition results depend only on the target's intrinsic attributes, free from irrelevant factors. Experiments on the aircraft electromagnetic simulation dataset and the measured dataset demonstrate that the proposed method achieves higher accuracy and stronger robustness in HRRP target recognition without requiring any aspect labels or prior knowledge of disturbances.

For unmanned aerial vehicle (UAV) ground crew marshalling tasks, the accuracy of skeleton-based action recognition is often limited by the high similarity of motion patterns across action categories as well as variations in individual performance. To address this issue, we propose an adaptive refined graph convolutional network with enhanced features for action recognition. First, a multi-order and motion feature modeling module is constructed, which integrates joint positions, skeletal structures, and angular encodings for multi-granularity representation. Static-domain and dynamic-domain features are then fused to enhance the diversity and expressiveness of the input representations. Second, a data-driven adaptive graph convolution module is designed, where inter-joint interactions are dynamically modeled through a learnable topology. Furthermore, an adaptive refinement feature activation mechanism is introduced to optimize information flow between nodes, enabling fine-grained modeling of skeletal spatial information. Finally, a frame-index semantic temporal modeling module is incorporated, where joint-type semantics and frame-index semantics are introduced in the spatial and temporal dimensions, respectively, to capture the temporal evolution of actions and comprehensively exploit spatio-temporal semantic correlations. On the NTU-RGB+D 60 and NTU-RGB+D 120 benchmark datasets, the proposed method achieves accuracies of 89.4% and 94.2% under X-Sub and X-View settings, respectively, as well as 81.7% and 83.3% on the respective benchmarks. On the self-constructed UAV Airfield Ground Crew Dataset, the proposed method attains accuracies of 90.71% and 96.09% under X-Sub and HO settings, respectively. Environmental robustness experiments demonstrate that under complex environmental conditions including illumination variations, haze, rain, shadows, and occlusions, the adoption of the Test + Train strategy reduces the maximum performance degradation from 3.1 percentage points to within 1 percentage point. Real-time performance testing shows that the system achieves an end-to-end inference latency of 24.5 ms (40.8 FPS) on the edge device NVIDIA Jetson Xavier NX, meeting real-time processing requirements and validating the efficiency and practicality of the proposed method on edge computing platforms.

Satellite segmentation is vital for spacecraft perception, supporting tasks like structural analysis, fault detection, and in-orbit servicing. However, the generalization of existing methods is severely limited by the scarcity of target satellite data and substantial morphological differences between target satellites and training samples, leading to suboptimal performance in real-world scenarios. In this work, we propose a novel diffusion-based framework for few-shot satellite segmentation, named DiffSatSeg, which leverages the powerful compositional generalization capability of diffusion models to address the challenges inherent in satellite segmentation tasks. Specifically, we propose a parameter-efficient fine-tuning strategy that fully exploits the strong prior knowledge of diffusion models while effectively accommodating the unique structural characteristics of satellites as rare targets. We further propose a segmentation mechanism based on distributional similarity, designed to overcome the limited generalization capability of conventional segmentation models when encountering novel satellite targets with substantial inter-class variations. Finally, we design a consistency learning strategy to suppress redundant texture details in diffusion features, thereby mitigating their interference in segmentation. Extensive experiments demonstrate that our method achieves state-of-the-art performance, yielding a remarkable 33.6% improvement over existing approaches even when only a single target satellite image is available. Notably, our framework also enables reference-based segmentation, which holds great potential for practical deployment and real-world applications.

Open-vocabulary semantic segmentation (OVSS) is of critical importance for unmanned aerial vehicle (UAV) imagery, as UAV scenes are highly dynamic and characterized by diverse, unpredictable object categories. Current OVSS approaches mainly rely on the zero-shot capabilities of vision–language models (VLMs), but their image-level pretraining objectives yield ambiguous spatial relationships and coarse-grained feature representations, resulting in suboptimal performance in UAV scenes. In this work, we propose a novel hybrid framework for OVSS in UAV imagery, named HOSU, which leverages the priors of vision foundation models to unleash the potential of vision–language models in representing complex spatial distributions and capturing fine-grained small-object details in UAV scenes. Specifically, we propose a distribution-aware fine-tuning method that aligns CLIP with DINOv2 across intra- and inter-region feature distributions, enhancing the capacity of CLIP to model complex scene semantics and capture fine-grained details critical for UAV imagery. Meanwhile, we propose a text-guided multi-level regularization mechanism that leverages the text embeddings of CLIP to impose semantic constraints on the visual features, preventing their drift from the original semantic space during fine-tuning and ensuring stable vision–language correspondence. Finally, to address the pervasive occlusion in UAV imagery, we propose a mask-based feature consistency strategy that enables the model to learn stable representations, remaining robust against viewpoint-induced occlusions. Extensive experiments across four training settings on six UAV datasets demonstrate that our approach consistently achieves state-of-the-art performance compared with previous methods, while comprehensive ablation studies and analyses further validate its effectiveness.