Due to the rapid development of deep learning, the performance of object detection has greatly improved. However, object detection in high-resolution Unmanned Aerial Vehicles images remains a challenging problem for three main reasons: (1) the objects in aerial images have different scales and are usually small; (2) the images are high-resolution but state-of-the-art object detection networks are of a fixed size; (3) the objects are not evenly distributed in aerial images. To this end, we propose a two-stage Adaptive Region Selection Detection framework in this paper. An Overall Region Detection Network is first applied to coarsely localize the object. A fixed points density-based targets clustering algorithm and an adaptive selection algorithm are then designed to select object-dense sub-regions. The object-dense sub-regions are sent to a Key Regions Detection Network where results are fused with the results at the first stage. Extensive experiments and comprehensive evaluations on the VisDrone2021-DET benchmark datasets demonstrate the effectiveness and adaptiveness of the proposed framework. Experimental results show that the proposed framework outperforms, in terms of mean average precision (mAP), the existing baseline methods by 2.1% without additional time consumption.

On 10 May 2024, a super geomagnetic storm with a minimum Dst index of less than −400 nT occurred. It has attracted a significant amount of attention in the literature. Using total electron content (TEC) observations from a global navigation satellite system (GNSS), in situ electron density data from the Swarm satellite, and corresponding simulations from the thermosphere–ionosphere–electrodynamics general circulation model (TIEGCM), the dynamic poleward shift of the equatorial ionization anomaly (EIA) during the main phase of the super geomagnetic storm has been explored. The results show that the EIA crests moved poleward from ±15° magnetic latitude (MLat) to ±20° MLat at around 19.6 UT, to ±25° MLat at 21.2 UT, and to ±31° MLat at 22.7 UT. This poleward shift was primarily driven by the enhanced eastward electric field, neutral winds, and ambipolar diffusion. Storm-induced meridional winds can move ionospheric plasma upward/downward along geomagnetic field lines, causing the poleward movement of EIA crests, with minor contributions from zonal winds. Ambipolar diffusion contributes/prevents the formation of EIA crests at most EIA latitudes/the equatorward edge.

Drones are extensively utilized in both military and social development processes. Eliminating the reliance of drone positioning systems on GNSS and enhancing the accuracy of the positioning systems is of significant research value. This paper presents a novel approach that employs a real-scene 3D model and image point cloud reconstruction technology for the autonomous positioning of drones and attains high positioning accuracy. Firstly, the real-scene 3D model constructed in this paper is segmented in accordance with the predetermined format to obtain the image dataset and the 3D point cloud dataset. Subsequently, real-time image capture is performed using the monocular camera mounted on the drone, followed by a preliminary position estimation conducted through image matching algorithms and subsequent 3D point cloud reconstruction utilizing the acquired images. Next, the corresponding real-scene 3D point cloud data within the point cloud dataset is extracted in accordance with the image-matching results. Finally, the point cloud data obtained through image reconstruction is matched with the 3D point cloud of the real scene, and the positioning coordinates of the drone are acquired by applying the pose estimation algorithm. The experimental results demonstrate that the proposed approach in this paper enables precise autonomous positioning of drones in complex urban environments, achieving a remarkable positioning accuracy of up to 0.4 m.

To address the challenge of collaborative nodes being unable to accurately perceive each other’s positions in global navigation satellite system (GNSS)-denied environments (such as after hostile interference or in urban canyons), we propose a GNSS-independent collaborative positioning radio frequency (RF) sensor. This sensor estimates inter-node distances and orientations using wireless measurements between nodes, without requiring pre-deployed anchor points. First, we designed a low-nanosecond latency ranging logic circuit on field-programmable gate array (FPGA) hardware, enabling relative distance estimation between nodes via a low-latency collaborative ranging (LLCR) algorithm without synchronization. Additionally, a synthetic aperture rotating antenna system was built to construct an echo space energy distribution matrix, based on dynamic–static dual-channel phase differences for high-precision, unambiguous azimuth measurement, followed by angle and distance data integration for localization. Then, a novel RF sensor hardware system was designed that was lightweight, low in cost, and high in performance. Finally, two generations of prototype models were developed and tested in both an anechoic chamber and mounted on unmanned vehicles outdoors in fields. The results demonstrate that the proposed sensor can achieve high-precision relative position estimation between collaborative nodes in the absence of GNSS, with a positioning error of within 0.4 m, indicating that it is suitable for mounting on unmanned vehicles and other autonomous systems for collaborative positioning.

Defects in photovoltaic (PV) modules, including hotspots, shading, and diode failures, significantly reduce power-generation efficiency and pose safety risks. This study proposes a real-time detection and localisation framework for PV defects based on infrared images acquired by unmanned aerial vehicles (UAVs). A dedicated dataset of 5583 infrared/visible images was constructed under standardised acquisition conditions. An improved rotating-bounding-box detector, termed YOLO-CLO, was developed upon YOLOv8-OBB by introducing a lightweight C3m module and a shared-convolution LSCD-OBB detection head. The proposed detector attains 99.1% mAP@0.5, 96.7% mAP@0.5:0.95, and 59.88 FPS with only 8.52 M parameters and 23.6 GFLOPs, outperforming the baseline in both accuracy and efficiency. A multi-feature image-processing pipeline combining gradient, grayscale, temperature, and morphological cues identifies hotspots, diode failures, and obstructions with detection accuracies of 96.97%, 100%, and 88.89%, respectively. A component-level localisation strategy integrating GNSS metadata, the Hough transform, and an improved K-means clustering algorithm accurately recovers the row–column index of each defective module within an array. Comparative experiments against YOLOv5 and Faster R-CNN confirm the superiority of the proposed framework. The method offers low hardware dependency and is suitable for engineering deployment in large-scale PV power stations.

Precise cooperative node localization is essential for the application of multifunctional integrated radio frequency (RF) sensor networks in military and civilian domains. Most geometric localization methods commonly rely on observation data from multiple receiving nodes or anchor points with known positions and synchronized clocks, producing complex system architectures and high construction costs. To address this, our paper proposes an effective single-station cooperative node localization technique, where the observation station only requires two antennas for operation. Leveraging prior knowledge of the geometry of surrounding structures, multiple virtual stations (VSs) are constructed by mining the spatiotemporal information contained in the multipath components (MPCs) to realize target positioning. The proposed method consists of two steps. In the first step, an unambiguous dual-antenna direction-finding algorithm is designed to extract the spatial information of MPCs and construct VSs, allowing a preliminary estimate of the source position (SP). In the second step, the path delays are extracted via matched filtering, while the spatiotemporal information is correlated based on the energy distribution for a more precise SP estimation. Simulations and experimental results demonstrate that our algorithm achieves high-precision single-station localization for a collaborative node, with positioning accuracy typically within $0.1$ m.