Detecting small objects in drone imagery is challenging due to low resolution and background blending, leading to limited feature information. Multi-scale feature fusion can enhance detection by capturing information at different scales, but traditional strategies fall short. Simple concatenation or addition operations do not fully utilize multi-scale fusion advantages, resulting in insufficient correlation between features. This inadequacy hinders the detection of small objects, especially in complex backgrounds and densely populated areas. To address this issue and efficiently utilize the limited computational resources, we propose a lightweight fusion strategy based on enhanced inter-layer feature correlation (EFC) to replace the traditional feature fusion strategy in FPN. The semantic expressions of different layers in the feature pyramid are inconsistent. In EFC, the Grouped Feature Focus Unit (GFF) enhances the feature correlation of each layer by focusing on the contextual information of different features. The Multi-Level Feature Reconstruction Module (MFR) effectively reconstructs and transforms the strength and weakness information of each layer in the pyramid to reduce redundant feature fusion and retain more information about small targets in deep networks. It is noteworthy that the proposed method is plug-and-play and can be widely applied to various base networks. Extensive experiments and comprehensive evaluations on VisDrone, UAVDT and COCO demonstrate the effectiveness. Using GFL as the baseline on the VisDrone dataset with a large number of small targets, the proposed method improves the detection mAP by 1.7%, surpassing many lightweight state-of-the-art methods and significantly reducing the Params and GFLOPs at the neck end.

In the realm of unmanned aerial vehicle (UAV) tracking, Siamese-based approaches have gained traction due to their optimal balance between efficiency and precision. However, UAV scenarios often present challenges such as insufficient sampling resolution, fast motion and small objects with limited feature information. As a result, temporal context in UAV tracking tasks plays a pivotal role in target location, overshadowing the target’s precise features. In this paper, we introduce MT-Track, a streamlined and efficient multi-step temporal modeling framework designed to harness the temporal context from historical frames for enhanced UAV tracking. This temporal integration occurs in two steps: correlation map generation and correlation map refinement. Specifically, we unveil a unique temporal correlation module that dynamically assesses the interplay between the template and search region features. This module leverages temporal information to refresh the template feature, yielding a more precise correlation map. Subsequently, we propose a mutual transformer module to refine the correlation maps of historical and current frames by modeling the temporal knowledge in the tracking sequence. This method significantly trims computational demands compared to the raw transformer. The compact yet potent nature of our tracking framework ensures commendable tracking outcomes, particularly in extended tracking scenarios. Comprehensive tests across four renowned UAV benchmarks substantiate the superior efficacy of our approach, delivering real-time performance at 84.7 FPS on a single GPU. Real-world test on the NVIDIA AGX hardware platform achieves a speed exceeding 30 FPS, validating the practicality of our method.

Detecting tiny objects in multimodal Red-Green-Blue-Thermal (RGBT) imagery is a critical challenge in computer vision, particularly in surveillance, search and rescue, and autonomous navigation. Drone-based scenarios exacerbate these challenges due to spatial misalignment, low-light conditions, occlusion, and cluttered backgrounds. Current methods struggle to leverage the complementary information between visible and thermal modalities effectively. We propose COXNet, a novel framework for RGBT tiny object detection, addressing these issues through three core innovations: i) the Cross-Layer Fusion Module, fusing high-level visible and low-level thermal features for enhanced semantic and spatial accuracy; ii) the Dynamic Alignment and Scale Refinement module, correcting cross-modal spatial misalignments and preserving multi-scale features; and iii) an optimized label assignment strategy using the GeoShape Similarity Measure for better localization. COXNet achieves a 3.32% mAP₅₀ improvement on the RGBTDronePerson dataset over state-of-the-art methods, demonstrating its effectiveness for robust detection in complex environments.

Rotating object detection in remote sensing image has made substantial progress. However, the widely used feature extraction and fusion modules cannot accurately represent the spatial location and angle information of remote sensing objects in arbitrary orientation. A detection scheme that accurately expresses spatial angle characteristics is necessary when remote sensing objects are densely distributed. This paper proposes an effective network for rotating object detection in optical remote sensing image (BDR-Net). Firstly, a feature aggregation module (FAM) is designed to refine the feature extraction of rotating objects, which can generate refined feature maps. Secondly, two-dimensional (2-D) Gaussian modeling simulates the rotation angle information of the object. Thirdly, given that the Bhattacharyya distance is mathematically symmetric and rigorous, we design a Bhattacharyya distance based learning module (BDLM). We improve the loss function to more accurately guide the regression of the relative location information of a rotating box. The effectiveness of the proposed method is validated by the results of experiments on three public datasets: DOTA, DIOR, and HRSC2016. The results of our research prove that a Bhattacharyya distance with a Gaussian distribution could be used as an approximate regression loss of the object detection model. This paper offers an original approach to combining object detection with metric learning to apply rotating object detection to instrumentation in a deployable and rapid manner.

The inverse synthetic aperture radar (ISAR) image is a kind of target feature data acquired by radar for moving targets, which can reflect the shape, structure, and motion information of the target, and has attracted a great deal of attention from the radar automatic target recognition (RATR) community. The identification of ISAR image components in radar satellite identification missions has not been carried out in related research, and the relevant segmentation methods of optical images applied to the research of semantic segmentation of ISAR images do not achieve ideal segmentation results. To address this problem, this paper proposes an ISAR image part recognition method based on semantic segmentation and mask matching. Furthermore, a reliable automatic ISAR image component labeling method is designed, and the satellite target component labeling ISAR image samples are obtained accurately and efficiently, and the satellite target component labeling ISAR image data set is obtained. On this basis, an ISAR image component recognition method based on semantic segmentation and mask matching is proposed in this paper. U-Net and Siamese Network are designed to complete the ISAR image binary semantic segmentation and binary mask matching, respectively. The component label of the ISAR image is predicted by the mask matching results. Experiments based on satellite component labeling ISAR image datasets confirm that the proposed method is feasible and effective, and it has greater comparative advantages compared to other classical semantic segmentation networks.