Cross-view geolocalization of unmanned aerial vehicles (UAVs) is a challenging task due to the positional discrepancies and uncertainties in scale and distance between UAVs and satellite views. Existing transformer-based geolocalization methods mainly use encoders to mine image contextual information. However, these methods have some limitations when dealing with scale changes in cross-view images. Therefore, we present an effective transformer-based Siamese network tailored for UAV geolocalization, called GeoFormer. First, an efficient transformer feature extraction network was designed, which utilizes linear attention to reduce the computational complexity and improve the computational efficiency of the network. Among them, we designed an efficient separable perceptron module based on depthwise separable convolution, which can effectively reduce the computational cost while improving the feature representation of the network. Second, we proposed a multiscale feature aggregation module, which deeply fuses salient features at different scales through a feedforward neural network to generate global feature representations with rich semantics, which improves the model's ability to capture image details and represent robust features. Additionally, we designed a semantic-guided region segmentation module, which utilizes a k -modes clustering algorithm to divide the feature map into multiple regions with semantic consistency and performs feature recognition within each semantic region to improve the accuracy of image matching. Finally, we designed a hierarchical reinforcement rotation matching strategy to achieve accurate UAV geolocalization based on the retrieval results of UAV view query satellite images using SuperPoint keypoints extraction and LightGlue rotation matching. According to the experimental results, our method effectively achieves UAV geolocalization.

In this article, we propose an effective infrared and visible image fusion network based on contrastive learning, which is called CLF-Net. A novel noise contrastive estimation framework is introduced into the image fusion to maximize mutual information between the fused image and source images. First, an unsupervised contrastive learning framework is constructed to promote fused image selectively retaining the most similar features in local areas of different source images. Second, we design a robust contrastive loss based on the deep representations of images, combining with the structural similarity loss to effectively guide the network in extracting and reconstructing features. Specifically, based on the deep representation similarities and structural similarities between the fused image and source images, the loss functions can guide the feature extraction network in adaptively obtaining the salient targets of infrared images and background textures of visible images. Then, the features are reconstructed in the most appropriate manner. In addition, our method is an unsupervised end-to-end model. All of our methods have been tested on public datasets. Based on extensive qualitative and quantitative analysis results, it has been demonstrated that our proposed method performs better than the existing state-of-the-art fusion methods.

The challenges in UAV visual geo-localization primarily stem from discrepancies between satellite maps and aerial images, including scale variations, viewpoint deviations, and spatiotemporal mismatches. Current approaches adopt retrieval-based or keypoint-matching-based localization, and some studies employ a cascaded approach. However, these methods still exhibit limitations in addressing discrepancies. To address these challenges, we propose a region-to-point UAV visual geo-localization framework named R2PLoc. Specifically, we consider UAV visual geo-localization as the process of retrieving corresponding regions from satellite map databases using aerial images while establishing projective relationship between them. First, we employ a shared backbone network for semantic feature extraction to conserve computational resources. Then, the Hierarchical Semantic Aggregation Module (HSAM) is designed to address the feature distribution shifts by fusing multi-scale semantics that combine both global contexts and local structures. Additionally, the Semantic-Enhanced Hierarchical Refinement Matcher (SHRM) is constructed to improve the geometric consistency of keypoint matching by integrating high-level semantic information. Furthermore, the UAV-R2P dataset is constructed for the region-to-point geo-localization task. The qualitative and quantitative experimental results demonstrate that our method outperforms most state-of-the-art methods with similar model size on most available datasets.

The abuse of UAVs poses a potential risk to social security, necessitating the investigation of anti-UAV methods to safeguard critical areas. However, the existing UAV countermeasures face challenges such as high environmental impact, restricted spatial deployment, and low cost-effectiveness. To address these limitations, we developed a novel anti-UAV system known as UAV Hunter, which adopts an airborne tether-net capture device with visual aids to counter unauthorized UAVs. This system employs an “Anti-UAV with UAV” scheme, comprising a ground control station and a net-capturing UAV. The operator utilizes the ground control station to determine the mission area and flight path and then controls the flight of the net-capturing UAV. During flight, the net-capturing UAV leverages its dual-mode sensor to continuously monitor the target area. Simultaneously, the onboard computer executes a UAV detection and tracking algorithm to search for unauthorized UAVs in real time. The results are relayed to the operator in real time, facilitating precise adjustments for the net-capturing UAV to launch the rope net accurately. The system successfully realizes the functions of dual-mode real-time detection and tracking, precise net capture, and efficient integrated control. Compared with existing methods, the developed system exhibits accurate recognition, rapid action, diverse application scenarios, and an enhanced human–machine interaction experience. Test results in the open environment further validate the feasibility and functional integrity of the system, demonstrating its capability to effectively capture low-altitude unauthorized UAVs.

It is well known that the diverse causes of low-light images challenge the adaptability of enhancement algorithms in uncertain environments. Most deep learning-based algorithms only learn single illuminance estimation or mapping relationship, which inhibit the generalization ability of the model. To address this, we propose a novel multi-illumination estimation framework based on ghost imaging theory, dubbed Ghillie. Specifically, we consider low-light enhancement as a re-imaging process for objects in dark scenes. First, the light modulation network (LMN) is designed to modulate a series of estimated lights following a normal light distribution. These lights "illuminate" the low-light image and the enhanced illuminance image can be reconstructed by a differential ghost imaging algorithm. Then, a gradient-guided denoising network (GDN) is constructed to eliminate noise and enhance details. Finally, we employ the color adaption network (CAN) to restore the color degradation. Additionally, a novel mean structural similarity loss (AM-SSIM) is proposed to guide the model to address the uneven image illumination. The qualitative and quantitative experimental results show that our enhanced methods outperform state-of-the-art methods on the vast majority of publicly available datasets.