Drone Target Research Articles

Multi-level representation learning via ConvNeXt-based network for unaligned cross-view matching

ABSTRACT Cross-view matching refers to the use of images from different platforms (e.g. drone and satellite views) to retrieve the most relevant images, where the key is that the viewpoints and spatial resolution. However, most of the existing methods focus on extracting fine-grained features and ignore the connection of contextual information in the image. Therefore, we propose a novel ConvNeXt-based multi-level representation learning model for the solution of this task. First, we extract global features through the ConvNeXt model. In order to obtain a joint part-based representation learning from the global features, we then replicated the obtained global features, operating one copy with spatial attention and the other copy using a standard convolutional operation. In addition, the features of different branches are aggregated through the multilevel feature fusion module to prepare for cross-view matching. Finally, we created a new hybrid loss function to better limit these features and assist in mining crucial data regarding global features. The experimental results indicate that we have achieved advanced performance on two common datasets, University-1652 and SUES-200 at 89.79% and 95.75% in drone target matching and 94.87% and 98.80 in drone navigation.

Edge-Guided Feature Pyramid Networks: An Edge-Guided Model for Enhanced Small Target Detection.

Infrared small target detection technology has been widely applied in the defense sector, including applications such as precision targeting, alert systems, and naval monitoring. However, due to the small size of their targets and the extended imaging distance, accurately detecting drone targets in complex infrared environments remains a considerable challenge. Detecting drone targets accurately in complex infrared environments poses a substantial challenge. This paper introduces a novel model that integrates edge characteristics with multi-scale feature fusion, named Edge-Guided Feature Pyramid Networks (EG-FPNs). This model aims to capture deep image features while simultaneously emphasizing edge characteristics. The goal is to resolve the problem of missing target information that occurs when Feature Pyramid Networks (FPNs) perform continuous down-sampling to obtain deeper semantic features. Firstly, an improved residual block structure is proposed, integrating multi-scale convolutional feature extraction and inter-channel attention mechanisms, with significant features being emphasized through channel recalibration. Then, a layered feature fusion module is introduced to strengthen the shallow details in images while fusing multi-scale image features, thereby strengthening the shallow edge features. Finally, an edge self-fusion module is proposed to enhance the model's depiction of image features by extracting edge information and integrating it with multi-scale features. We conducted comparative experiments on multiple datasets using the proposed algorithm and existing advanced methods. The results show improvements in the IoU, nIoU, and F1 metrics, while also showcasing the lightweight nature of EG-FPNs, confirming that they are more suitable for drone detection in resource-constrained infrared scenarios.

YOLOv8-MDS: A YOLOv8-Based Multi-Distance Scale Drone Detection Network

Abstract Drones have become widely used across various fields, showcasing their capabilities while also raising significant security and privacy concerns. Current detection methods, such as radar, radio frequency, and acoustic detection systems, face issues like high costs and poor interference resistance. The rapid development of computer vision has led to the emergence of visual-based drone detection solutions. To address the challenge of varying drone sizes at different distances impacting visual detection, this paper constructs a multi-distance scale drone dataset with images captured at different distances and environments. Additionally, a Multi-Distance Scale Feature Attention Module (MDS-Module) is proposed and integrated into the neck of the model via leap connections to enhance global feature detection. Furthermore, to account for the typically rectangular nature of drone anchor boxes, the EIoU loss function is used in the detection head to improve the model’s detection capability for drone targets. we conducted comprehensive ablation and comparative experiments on the improved model. The experimental results demonstrate that incorporating the Multi-Distance Scale Feature Attention Module significantly enhances the model’s ability to detect drone targets across multiple distance scales.

A Vision-Based End-to-End Reinforcement Learning Framework for Drone Target Tracking

A Vision-Based End-to-End Reinforcement Learning Framework for Drone Target Tracking

Real-Time Ubiquitous Radar Target Classification with 1D ResNet-SE-Based Multi-Channel Network

Ubiquitous radar has significant advantages over traditional radar in detecting and identifying low, slow, and small (LSS) targets in a strong clutter environment. It effectively addresses challenges faced in low-altitude target monitoring within the low-altitude economy (LAE). The working mode of ubiquitous radar, which tracks first and identifies later, provides high-resolution Doppler data to the target identification module. Utilizing high-resolution Doppler data allows for the effective identification of LSS targets. To meet the needs of real-time classification, this paper first designs a real-time classification process based on sliding window Doppler data. This process requires the classifier to classify targets based on multiple rows of high-resolution Doppler spectra within the sliding window. Secondly, a multi-channel parallel perception network based on a 1D ResNet-SE network is designed. This network captures features within the rows of sliding window data and integrates inter-row features. Experiments show that the designed real-time classification process and multi-channel parallel perception network meet real-time classification requirements. Compared to the 1D CNN-MLP multi-channel network, the proposed 1D ResNet-MLP multi-channel network improves the classification accuracy from 98.71% to 99.34%. Integrating the 1D Squeeze-and-Excitation (SE) module to form the 1D ResNet-SE-MLP network further enhances accuracy to 99.58%, with drone target accuracy, the primary focus of the LAE, increasing from 97.19% to 99.44%.

A video speed measurement method for UAV targeting ground targets based on kalman filtering

Abstract Measuring the velocity information of moving targets, as an application scenario for drone target velocity measurement, is an important focus in the process of drone intelligence reconnaissance. To improve the speed and accuracy of unmanned aerial vehicles (UAVs) in locating and measuring moving targets, a reliable target localization and velocity model based on UAV reconnaissance video frames and Kalman filtering is proposed. This algorithm model can complete the localization and velocity measurement tasks of moving targets under complex conditions without the need to obtain ranging information from drones. To verify the algorithm, the Monte Carlo simulation method was used for simulation experiments. The experimental results show that the error of this velocity measurement algorithm can meet the accuracy requirements of unmanned aerial vehicles for ground-moving targets.

LD-YOLOv10: A Lightweight Target Detection Algorithm for Drone Scenarios Based on YOLOv10

Due to the limited computing resources and storage capacity of edge detection devices, efficient detection algorithms are typically required to meet real-time and accuracy requirements. Existing detectors often require a large number of parameters and high computational power to improve accuracy, which reduces detection speed and performance on low-power devices. To reduce computational load and enhance detection performance on edge devices, we propose a lightweight drone target detection algorithm, LD-YOLOv10. Firstly, we design a novel lightweight feature extraction structure called RGELAN, which utilizes re-parameterized convolutions and the newly designed Conv-Tiny as the computational structure to reduce the computational burden of feature extraction. The AIFI module was introduced, utilizing its multi-head attention mechanism to enhance the expression of semantic information. We construct the DR-PAN Neck structure, which obtains weak features of small targets with minimal computational load. Wise-IoU and EIoU are combined as new bounding box regression loss functions to adjust the competition between anchor boxes of different quality and the sensitivity of anchor box aspect ratios, providing a more intelligent gradient allocation strategy. Extensive experiments on the VisdroneDET-2021 and UAVDT datasets show that LD-YOLOv10 reduces the number of parameters by 62.4% while achieving a slight increase in accuracy and has a faster detection speed compared to other lightweight algorithms. When deployed on the low-power NVIDIA Jetson Orin Nano device, LD-YOLOv10 achieves a detection speed of 25 FPS.

Simulation Study of the Damage Effect of Typical Fragment on UAV Target

Abstract UAV poses a serious threat to ground troops and combat personnel in modern operations, and fighting against UAV is called an important task of air defense operations. Taking DJI UAV as the object, this article researches the typical blasting debris of the prefabricated damage effect of drone target through modeling simulation and simulation calculation, which is focused on damage simulation analysis of the drone battery, power plant, camera and other key components, and studies the damage condition of fragments in different direction and at different speeds, which can improve theoretical basis and reference for fighting against the drone and its swarm.

A More Efficient Algorithm for Small Target Detection in Unmanned Aerial Vehicles

Due to the relatively high shooting altitude of unmanned aerial vehicles (UAV), the captured images often contain a multitude of small‐scale targets. To solve the problems of small target scale, lack of semantic information, and high miss detection in drone target detection, in this paper we proposed a more effective unmanned aerial vehicle small target detection algorithm(MEU‐YOLOv5) based on YOLOv5s. Firstly, an efficient global contextual module is proposed to enhance the algorithm's performance in feature extraction while reducing the excessive loss of shallow features. Secondly, a small‐scale target detector is added to enhance the algorithm's detection capability for smaller targets. Lastly, a recursive multi‐level feature fusion path is introduced to better fuse the shallow and deep features of the images, reducing overfitting and improving the algorithm's generalizability and robustness. Experimental results demonstrated that compared to YOLOv5s, MEU‐YOLOv5 achieves a 7.4% improvement in mAP@0.5 and a 4.9% improvement in mAP@0.5:0.95. Additionally, the overall performance of this algorithm surpassed various algorithms in the YOLO series, including YOLOv3, YOLOv5l, YOLOv5m, and YOLOv8s. © 2024 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.

Lightweight Saliency Target Intelligent Detection based on Multi-scale Feature Adaptive Fusion

In order to solve the problems of small targets, variable shooting angles, and heights in drone images, the author proposes an adaptive drone target intelligent detection algorithm based on multi-scale feature fusion. The results show that after adding a deconvolution cascade structure to the network, mAP increased by about 2.5 percentage points and AP$^{50}$ increased by about 3 percentage points. Compared with Method 3, Method 4 uses GA-RPN instead of RPN, and when the IOU is 75, the AP increases by 3.5 percentage points, reflecting that the target prediction candidate boxes generated using semantic features adaptively match better than the manually designed target candidate boxes. This indicates that the proposed target detection framework has better classification ability and higher frame regression accuracy. Multi scale adaptive candidate regions are used to generate fused features of different scales generated by the network, weighted fused multi-scale features are used for target prediction, and semantic features are used to guide the network to adaptively generate target candidate frames, greatly enhancing the feature expression ability of various targets and improving the detection accuracy of aerial targets.

Lightweight air-to-air unmanned aerial vehicle target detection model

The rapid expansion of the drone industry has resulted in a substantial increase in the number of low-altitude drones, giving rise to concerns regarding collision avoidance and countermeasure strategies among these unmanned aerial vehicles. These challenges underscore the urgent need for air-to-air drone target detection. An effective target detection model must exhibit high accuracy, real-time capabilities, and a lightweight network architecture to achieve a balance between precision and speed when deployed on embedded devices. In response to these requirements, we initially curated a dataset comprising over 10,000 images of low-altitude operating drones. This dataset encompasses diverse and intricate backgrounds, significantly enhancing the model’s training capacity. Subsequently, a series of enhancements were applied to the YOLOv5 algorithm to realize lightweight object detection. A novel feature extraction network, CF2-MC, streamlined the feature extraction process, while an innovative module, MG, in the feature fusion section aimed to improve detection accuracy and reduce model complexity. Concurrently, the original CIoU loss function was replaced with the EIoU loss function to further augment the model’s accuracy. Experimental results demonstrate an enhancement in the accuracy of drone target detection, achieving mAP values of 95.4% on the UAVfly dataset and 82.2% on the Det-Fly dataset. Finally, real-world testing conducted on the Jetson TX2 revealed that the YOLOv5s-ngn model achieved an average inference speed of 14.5 milliseconds per image. The code utilized in this paper can be accessed via https://github.com/lucien22588/yolov5-ngn.git.

Simultaneous Tracking and Recognizing Drone Targets with Millimeter-Wave Radar and Convolutional Neural Network

In this paper, a framework for simultaneous tracking and recognizing drone targets using a low-cost and small-sized millimeter-wave radar is presented. The radar collects the reflected signals of multiple targets in the field of view, including drone and non-drone targets. The analysis of the received signals allows multiple targets to be distinguished because of their different reflection patterns. The proposed framework consists of four processes: signal processing, cloud point clustering, target tracking, and target recognition. Signal processing translates the raw collected signals into spare cloud points. These points are merged into several clusters, each representing a single target in three-dimensional space. Target tracking estimates the new location of each detected target. A novel convolutional neural network model was designed to extract and recognize the features of drone and non-drone targets. For the performance evaluation, a dataset collected with an IWR6843ISK mmWave sensor by Texas Instruments was used for training and testing the convolutional neural network. The proposed recognition model achieved accuracies of 98.4% and 98.1% for one and two targets, respectively.

Drone Detection and Tracking System Based on Fused Acoustical and Optical Approaches

The increasing popularity of small drones has stressed the urgent need for an effective drone‐oriented surveillance system that can work day and night. Herein, an acoustic and optical sensor‐fusion‐based system‐termed multimodal unmanned aerial vehicle 3D trajectory exposure system (MUTES) is presented to detect and track drone targets. MUTES combines multiple sensor modules including microphone array, camera, and lidar. The 64‐channel microphone array provides semispherical surveillance with high signal‐to‐noise ratio of sound source estimation, while the long‐range lidar and the telephoto camera are capable of subsequent precise target localization in a narrower but higher definition field of view. MUTES employs a coarse‐to‐fine, passive‐to‐active localization strategy for wide‐range detection (semispherical) and high‐precision 3D tracking. To further increase the fidelity, an environmental denoising model is trained, which helps to select valid acoustic features from a drone target, thus overcomes the drawbacks of the traditional sound source localization approaches when facing noise interference. The effectiveness of the proposed sensor‐fusion approach is validated through field experiments. To the best of the knowledge, MUTES provides the farthest detection range, highest 3D position accuracy, strong anti‐interference capability, and acceptable cost for unverified drone intruders.

Enhancing UAV Detection in Surveillance Camera Videos through Spatiotemporal Information and Optical Flow.

The growing intelligence and prevalence of drones have led to an increase in their disorderly and illicit usage, posing substantial risks to aviation and public safety. This paper focuses on addressing the issue of drone detection through surveillance cameras. Drone targets in images possess distinctive characteristics, including small size, weak energy, low contrast, and limited and varying features, rendering precise detection a challenging task. To overcome these challenges, we propose a novel detection method that extends the input of YOLOv5s to a continuous sequence of images and inter-frame optical flow, emulating the visual mechanisms employed by humans. By incorporating the image sequence as input, our model can leverage both temporal and spatial information, extracting more features of small and weak targets through the integration of spatiotemporal data. This integration augments the accuracy and robustness of drone detection. Furthermore, the inclusion of optical flow enables the model to directly perceive the motion information of drone targets across consecutive frames, enhancing its ability to extract and utilize features from dynamic objects. Comparative experiments demonstrate that our proposed method of extended input significantly enhances the network's capability to detect small moving targets, showcasing competitive performance in terms of accuracy and speed. Specifically, our method achieves a final average precision of 86.87%, representing a noteworthy 11.49% improvement over the baseline, and the speed remains above 30 frames per second. Additionally, our approach is adaptable to other detection models with different backbones, providing valuable insights for domains such as Urban Air Mobility and autonomous driving.

PFFNET: A Fast Progressive Feature Fusion Network for Detecting Drones in Infrared Images

The rampant misuse of drones poses a serious threat to national security and human life. Currently, CNN (Convolutional Neural Networks) are widely used to detect drones. However, small drone targets often reduced amplitude or even lost features in infrared images which traditional CNN cannot overcome. This paper proposes a Progressive Feature Fusion Network (PFFNET) and designs a Pooling Pyramid Fusion (PFM) to provide more effective global contextual priors for the highest downsampling output. Then, the Feature Selection Model (FSM) is designed to improve the use of the output coding graph and enhance the feature representation of the target in the network. Finally, a lightweight segmentation head is designed to achieve progressive feature fusion with multi-layer outputs. Experimental results show that the proposed algorithm has good real-time performance and high accuracy in drone target detection. On the public dataset, the intersection over union (IOU) is improved by 2.5% and the detection time is reduced by 81%.

Use of Plant Health Level Based on Random Forest Algorithm for Agricultural Drone Target Points

Chemical residues from the use of pesticides in agriculture can impact human health through environmental and food pollution. To lessen the negative effects of excessive pesticide use, pesticides must be applied to plants by dose. The dose of pesticide application can be based on a plant health level, which is the result of drone Normalized Difference Vegetation Index (NDVI) image analysis. Drones can also be used for spraying pesticides. Analysis of plant health levels was carried out using the Random Forest (RF) algorithm. The results of the classification plant health levels will be used to design spray drone flight routes. The objective of this research is to classify plant health levels of rice based on NDVI imagery using the RF algorithm and to compile a database of spray drone target points. The results of this study indicate that the classification of plant health levels using the RF algorithm produces an accuracy value of 98% and a Kappa value of 0.96. As a result, the model developed and the algorithm employed is quite effective at classifying the level of plant health. Furthermore, spray drone target points based on plant health levels can be generated. Optimally the spray distance between rows is 2 m.

The Notion of the “Subaltern” and the Drone Victim Subjectivities in Pakistani Anglophone Fiction

This research proposes a new approach to study fictional representation of drone victims in contemporary Anglophone Pakistani novels by examining the inevitable onto-epistemological relation between the notion of the “subaltern” and its possible implications for fictional drone subjectivities. By drawing insights from the newly emerging drone fiction literary theory and re-examining the already well-established tradition of Subaltern Studies, I demonstrate in this essay that the drone fiction written by Pakistani authors is a distinct literary category which configures drone victims as voiceless. I argue that by incorporating the impunity, inhumanity, and injustice inherent in the perpetration of drone warfare, Pakistani authors reimagine the drone victims as mute spectators of their devastation. The essay considers the narrative implications of drone technology’s orientalist, racist, and imperialist rubric in the selected works. A critical analysis of the selected textual excerpts unpacks the socio-political discursive subaltern formations of the fictional drone victim subjectivities such as dead bodies, widows, orphans, and disabled individuals. By unraveling the multiple layers of subjugation and repression embedded into the fictional representations of the drone victims, this essay proposes a possibility to study the precariousness of their situation under the praxis of Subaltern Studies. Moreover, different contexts and types of drone victims unravel the harsh conditions of the groups living under drone surveillance. The motif of drone target strikes allows Pakistani writers to explore the voicelessness of the dronized communities. Hence, in reading drone sufferers as subalterns (a subjugated mute political non-entity), the essay highlights the predicament of the marginalized dronized groups and foregrounds the possible expansion of Subaltern Studies in connection with the exploration of drone victim subjectivities.

Target Tracking based on Kernelized Correlation Filter Using MWIR and SWIR Sensors

When tracking small UAVs and drone targets in cloud clutter environments, MWIR sensors are often unable to track targets continuously. To overcome this problem, the SWIR sensor is mounted on the same gimbal. Target tracking uses sensor information fusion or selectively applies information from each sensor. In this case, parallax correction using the target distance is often used. However, it is difficult to apply the existing method to small UAVs and drone targets because the laser rangefinder's beam divergence angle is small, making it difficult to measure the distance. We propose a tracking method which needs not parallax correction of sensors. In the method, images from MWIR and SWIR sensors are captured simultaneously and a tracking error for gimbal driving is chosen by effectiveness measure. In order to prove the method, tracking performance was demonstrated for UAVs and drone targets in the real sky background using MWIR and SWIR image sensors.

Drone Detection Method Based on MobileViT and CA-PANet

Aiming at the problems of the large amount of model parameters and false and missing detections of multi-scale drone targets, we present a novel drone detection method, YOLOv4-MCA, based on the lightweight MobileViT and Coordinate Attention. The proposed approach is improved according to the framework of YOLOv4. Firstly, we use an improved lightweight MobileViT as the feature extraction backbone network, which can fully extract the local and global feature representations of the object and reduce the model’s complexity. Secondly, we adopt Coordinate Attention to improve PANet and to obtain a multi-scale attention called CA-PANet, which can obtain more positional information and promote the fusion of information with low- and high-dimensional features. Thirdly, we utilize the improved K-means++ method to optimize the object anchor box and improve the detection efficiency. At last, we construct a drone dataset and conduct a performance experiment based on the Mosaic data augmentation method. The experimental results show that the mAP of the proposed approach reaches 92.81%, the FPS reaches 40 f/s, and the number of parameters is only 13.47 M, which is better than mainstream algorithms and achieves a high detection accuracy for multi-scale drone targets using a low number of parameters.

A-RetinaNet: A novel RetinaNet with an asymmetric attention fusion mechanism for dim and small drone detection in infrared images.

To solve the problems of texture lacking and resolution coarseness in the detection of dim and small drone targets in infrared images, we propose a novel RetinaNet with an asymmetric attention fusion mechanism for dim and small drone detection. First, we propose a super-resolution texture-enhancement network as an effective solution for the lack of texture-related information on small infrared targets. The network generates super-resolution images and enhances the texture features of the targets. Second, considering the inadequacy of feature pyramids in the feature fusion stage, we use an asymmetric attention fusion mechanism to constitute an asymmetric attention fusion pyramid network for cross-layer feature fusion in a bidirectional manner; it achieves high-quality semantic and location detail information interaction between scale features. Third, a global average pooling layer is employed to capture global spatial-sensitive information, thus effectively identifying features and achieving classification. Experiments were conducted by using a publicly available infrared image dim-small drone target detection dataset; the results show that the proposed method achieves an AP of 95.43% and a recall of 80.6%, which is a significant improvement over the current mainstream target detection algorithms.