UAV Detection Research Articles

Feasibility Analysis of Optical UAV Detection Over Long Distances Using Robotic Telescopes

Substantial technological development has made Unmanned Aerial Vehicles (UAVs) more versatile, cheaper and accessible to the public in recent years. Alongside many positive effects and use cases, safety concerns are increasing as a plethora of incidents demonstrate the destructive potential of UAVs. To counteract this development and thus protect people and critical infrastructure, UAV detection, tracking and defence gains more and more research attention. Whereas, different drone detection technologies like RADAR, radio frequency and acoustic detection are deployed within multi-spectral systems, optical detection and imaging of approaching objects provide key information to correctly assess the situation. As reaction time is a crucial parameter for successful UAV defence, the operating distance of the optical detection system needs to be improved further. This paper presents the analysis, development and evaluation of a telescope-based UAV detection system. The system consists of a high precision mount and a telescope equipped with a camera. UAVs are detected in the captured video frames by the deep learning algorithm YOLOv4 using a modified architecture. The proposed system, which uses a f/10 telescope with a focal length of f = 2540 mm and a camera equipped with a 7.3 mm x 4.1 mm sensor, allows a significant increase of the optical detection range to more than 3 km of UAVs down to 0.3 m in diameter under daylight conditions and sufficient contrast, extending the reaction time significantly for counter UAV systems.

Mobility State Detection of Cellular-Connected UAVs Based on Handover Count Statistics

Mobility State Detection of Cellular-Connected UAVs Based on Handover Count Statistics

Estimation of Micro-Doppler Parameters With Combined Null Space Pursuit Methods for the Identification of LSS UAVs

Micro-Doppler (m-D) effect, induced by the rotation of rotor blades, supplies a differentiable characteristic to address the problem for the identification of low, slow and small unmanned aerial vehicles (LSS UAVs). However, the primary challenge for the estimation of m-D parameters is how to separate weak rotation signal from Doppler signal and other interferences. Theoretically, null space pursuit (NSP) is an operator-based signal decomposition approach to decompose a signal into additive subcomponents. The premise of NSP is that two separated components are orthogonal. However, due to the different modulation models, rotation signal, Doppler signal or other interferences could not satisfy the condition. Moreover, traditional multi-order differential operator is not suitable for the decomposition of m-D signal. In thi <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> s paper, back projection strategy with instantaneous orthogonal NSP (BPIO-NSP) is proposed to distill Doppler signal, and then micro-Doppler NSP (MD-NSP) is jointly developed to separate rotation signal for the identification of LSS UAVs. Firstly, the decomposed component after NSP is applied to the short-time Fourier transform (STFT) to find the segments with instantaneous non-orthogonal property. Secondly, the back projection strategy is developed in BPIO-NSP to acquire the instantaneous orthogonal data, so as to adjust the decomposed Doppler signal with high accuracy. Finally, customized operator is specially constructed in MD-NSP to achieve the required rotation signal from the residue. Simulation results verify the theoretical analysis, and measured data of the fun and UAV detection experiments suggest that the proposed methods could be served for the application of the identification of LSS UAVs.

Algorithm for estimating the energy distribution of radar signals scattering on acoustic disturbances created by UAVs

The task of estimating the energy distribution over the observation interval of radar signals scattered on atmospheric inhomogeneities, arising as a result of UAV operation, is considered. The solution to this problem is necessary to improve detection algorithms, to classify the detected UAVs according to additional informational features, to improve the resolution when detecting several devices located at the same range during the group application of UAVs, to clarify the time parameters of the evolution of the movement of UAVs in time and space. A similar problem arises due to the need to process useful radar signals with a low signal-to-noise ratio in order to achieve the maximum possible range of reliable UAV detection. Because of this, it becomes impossible to estimate directly the energy of useful signals by the method of comparison with reference physical quantities due to a large measurement error. Therefore, an evaluation algorithm is proposed, based on the methods of the theory of ordinal statistics, which provide, instead of comparing numerical realizations with a certain standard, to form a variational series from them under the condition of a priori knowledge of the distribution function of these realizations. At the same time, the fact is used that for certain distributions of a random variable, among which there are normal and all limited ones, the variance of the estimate in the form of a mathematical expectation of certain ordinal statistics is significantly less than the variance of a direct measurement at a low signal-to-noise ratio. In order to save time and computing resources during real-time processing of received signals, it is proposed to use pre-calculated arrays of numerical values of mathematical expectation and dispersion of ordinal statistics for various parameters of the density distribution of a random variable.

A Real-Time Tracking Algorithm for Multi-Target UAV Based on Deep Learning

UAV technology is a basic technology aiming to help realize smart living and the construction of smart cities. Its vigorous development in recent years has also increased the presence of unmanned aerial vehicles (UAVs) in people’s lives, and it has been increasingly used in logistics, transportation, photography and other fields. However, the rise in the number of drones has also put pressure on city regulation. Using traditional methods to monitor small objects flying slowly at low altitudes would be costly and ineffective. This study proposed a real-time UAV tracking scheme that uses the 5G network to transmit UAV monitoring images to the cloud and adopted a machine learning algorithm to detect and track multiple targets. Aiming at the difficulties in UAV detection and tracking, we optimized the network structure of the target detector yolo4 (You Only Look Once V4) and improved the target tracker DeepSORT, adopting the detection-tracking mode. In order to verify the reliability of the algorithm, we built a data set containing 3200 pictures of four UAVs in different environments, conducted training and testing on the model, and achieved 94.35% tracking accuracy and 69FPS detection speed under the GPU environment. The model was then deployed on ZCU104 to prove the feasibility of the scheme.

Deep-Ensemble-Learning-Based GPS Spoofing Detection for Cellular-Connected UAVs

Unmanned aerial vehicles (UAVs) are an emerging technology in the 5G-and-beyond systems with the promise of assisting cellular communications and supporting IoT deployment in remote and density areas. Safe and secure navigation is essential for UAV remote and autonomous deployment. Indeed, the opensource simulator can use commercial software-defined radio tools to generate fake global positioning system (GPS) signals and spoof the UAV GPS receiver to calculate wrong locations, deviating from the planned trajectory. Fortunately, the existing mobile positioning system can provide additional navigation for cellular-connected UAVs and verify the UAV GPS locations for spoofing detection, but it needs at least three base stations (BSs) at the same time. In this article, we propose a novel deep-ensemble-learning-based, mobile-network-assisted UAV monitoring and tracking system for cellular-connected UAV spoofing detection. The proposed method uses path losses between BSs and UAVs communication to indicate the UAV trajectory deviation caused by GPS spoofing. To increase the detection accuracy, three statistics methods are adopted to remove environmental impacts on path losses. In addition, deep ensemble learning methods are deployed on the edge cloud servers and use the multilayer perceptron (MLP) neural networks to analyze path losses statistical features for making a final decision, which has no additional requirements and energy consumption on UAVs. The experimental results show the effectiveness of our method in detecting GPS spoofing, achieving above 97% accuracy rate under two BSs, while it can still achieve at least 83% accuracy under only one BS.

A Comparison of Deep Neural Networks for Monocular Depth Map Estimation in Natural Environments Flying at Low Altitude

Currently, the use of Unmanned Aerial Vehicles (UAVs) in natural and complex environments has been increasing, because they are appropriate and affordable solutions to support different tasks such as rescue, forestry, and agriculture by collecting and analyzing high-resolution monocular images. Autonomous navigation at low altitudes is an important area of research, as it would allow monitoring parts of the crop that are occluded by their foliage or by other plants. This task is difficult due to the large number of obstacles that might be encountered in the drone's path. The generation of high-quality depth maps is an alternative for providing real-time obstacle detection and collision avoidance for autonomous UAVs. In this paper, we present a comparative analysis of four supervised learning deep neural networks and a combination of two for monocular depth map estimation considering images captured at low altitudes in simulated natural environments. Our results show that the Boosting Monocular network is the best performing in terms of depth map accuracy because of its capability to process the same image at different scales to avoid loss of fine details.

Effect of Lidar Receiver Field of View on UAV Detection

Researchers have shown that single-photon light detection and ranging (lidar) is highly sensitive and has a high temporal resolution. Due to the excellent beam directivity of lidar, most applications focus on ranging and imaging. Here, we present a lidar detection system for night environments. Different from MEMS, we choose a large divergence rather than scanning to detect unmanned aerial vehicles (UAVs). Collection and detection are achieved through the use of high-efficiency optical devices. With time-correlated single photon counting (TCSPC), we performed subsequent drone search work at centimeter resolution. We believe that we have developed a new technique for detecting UAVs. We show how the field of view influences the detection process. For some key areas of air defense, it is extremely necessary to find UAVs quickly and in a timely manner. In short, the results represent an important step toward practical, low-power drone detection using lidar.

Real-Time Detection of Drones Using Channel and Layer Pruning, Based on the YOLOv3-SPP3 Deep Learning Algorithm.

Achieving a real-time and accurate detection of drones in natural environments is essential for the interception of drones intruding into high-security areas. However, a rapid and accurate detection of drones is difficult because of their small size and fast speed. In this paper a drone detection method as proposed by pruning the convolutional channel and residual structures of YOLOv3-SPP3. First, the k-means algorithm was used to cluster label the boxes. Second, the channel and shortcut layer pruning algorithm was used to prune the model. Third, the model was fine tuned to achieve a real-time detection of drones. The experimental results obtained by using the Ubuntu server under the Python 3.6 environment show that the YOLOv3-SPP3 algorithm is better than YOLOV3, Tiny-YOLOv3, CenterNet, SSD300, and faster R-CNN. There is significant compression in the size, the maximum compression factor is 20.1 times, the maximum detection speed is increased by 10.2 times, the maximum map value is increased by 15.2%, and the maximum precision is increased by 16.54%. The proposed algorithm achieves the mAP score of 95.15% and the detection speed of 112 f/s, which can meet the requirements of the real-time detection of UAVs.

Drone and Controller Detection and Localization: Trends and Challenges

Unmanned aerial vehicles (UAVs) have emerged as a rapidly growing technology seeing unprecedented adoption in various application sectors due to their viability and low cost. However, UAVs have also been used to perform illegal and malicious actions, which have recently increased. This creates a need for technologies capable of detecting, classifying, and deactivating malicious and unauthorized drones. This paper reviews the trends and challenges of the most recent UAV detection methods, i.e., radio frequency-based (RF), radar, acoustic, and electro-optical, and localization methods. Our research covers different kinds of drones with a major focus on multirotors. The paper also highlights the features and limitations of the UAV detection systems and briefly surveys the UAV remote controller detection methods.

Robust UAV detection based on saliency cues and magnified features on thermal images

Robust UAV detection based on saliency cues and magnified features on thermal images

Research on Attitude Detection and Flight Experiment of Coaxial Twin-Rotor UAV.

Aiming at the problem that the single sensor of the coaxial UAV cannot accurately measure attitude information, a pose estimation algorithm based on unscented Kalman filter information fusion is proposed. The kinematics and dynamics characteristics of coaxial folding twin-rotor UAV are studied, and a mathematical model is established. The common attitude estimation methods are analyzed, and the extended Kalman filter algorithm and unscented Kalman filter algorithm are established. In order to complete the test of the prototype of a small coaxial twin-rotor UAV, a test platform for the dynamic performance and attitude angle of the semi-physical flight of the UAV was established. The platform can analyze the mechanical vibration, attitude angle and noise of the aircraft. It can also test and analyze the characteristics of the mechanical vibration and noise produced by the UAV at different rotor speeds. Furthermore, the static and time-varying trends of the pitch angle and yaw angle of the Kalman filter attitude estimation algorithm is further analyzed through static and dynamic experiments. The analysis results show that the attitude estimation of the UKF is better than that of the EKF when the throttle is between 0.2σ and 0.9σ. The error of the algorithm is less than 0.6°. The experiment and analysis provide a reference for the optimization of the control parameters and flight control strategy of the coaxial folding dual-rotor aircraft.

CUDM: A Combined UAV Detection Model Based on Video Abnormal Behavior

The widespread use of unmanned aerial vehicles (UAVs) has brought many benefits, particularly for military and civil applications. For example, UAVs can be used in communication, ecological surveys, agriculture, and logistics to improve efficiency and reduce the required workforce. However, the malicious use of UAVs can significantly endanger public safety and pose many challenges to society. Therefore, detecting malicious UAVs is an important and urgent issue that needs to be addressed. In this study, a combined UAV detection model (CUDM) based on analyzing video abnormal behavior is proposed. CUDM uses abnormal behavior detection models to improve the traditional object detection process. The work of CUDM can be divided into two stages. In the first stage, our model cuts the video into images and uses the abnormal behavior detection model to remove a large number of useless images, improving the efficiency and real-time detection of suspicious targets. In the second stage, CUDM works to identify whether the suspicious target is a UAV or not. Besides, CUDM relies only on ordinary equipment such as surveillance cameras, avoiding the use of expensive equipment such as radars. A self-made UAV dataset was constructed to verify the reliability of CUDM. The results show that CUDM not only maintains the same accuracy as state-of-the-art object detection models but also reduces the workload by 32%. Moreover, it can detect malicious UAVs in real-time.

A Non-Rigid Hierarchical Discrete Grid Structure and Its Application to UAVs Conflict Detection and Path Planning

<p xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Following the development of the unmanned aerial vehicle (UAV) industry, the number of UAVs has increased significantly, and higher expectations regarding the effect of spatial path planning have been expressed. Traditional coordinate-based conflict detection and path planning methods of UAVs are highly complex and have low computational efficiency. While the conflict detection method based on the grid system can improve efficiency, the rigid grid structure of the existing local grid system or discrete global grid system is insufficient. Since the rigid structure only encodes grid cells, multiple or different size grids are required to identify UAVs in motion, further complicating path planning. The main contributions of this article are as follows. First, we proposed a nonrigid hierarchical discrete grid structure and coding method for spatial three-dimensional grids and grid data model. This method can accomplish the aggregation of any adjacent eight grids to form multiple nonrigid structures, making the original fixed relationship between parent and child grids more flexible. Second, the nonrigid grid structure optimizes the identification ability of grid vertices, grid edges, and grid faces. It has eight times the identification ability of the rigid grid structure. We exploited this structure for UAV identification with the same size grid at any position of the grid structure. Third, based on the nonrigid grid structure, an improved UAV conflict detection method and path planning method are designed. Experimental results indicate that, on an average, the efficiency of the improved conflict detection is an order of magnitude higher than that of the traditional computational detection method. The improved path planning method reduces the path length by an average of 11% compared to the single-grid planning method, and is 3.85 times more efficient than the multi-grid buffer planning method.

Vision-Based Anti-UAV Detection and Tracking

Unmanned aerial vehicles (UAV) have been widely used in various fields, and their invasion of security and privacy has aroused social concern. Several detection and tracking systems for UAVs have been introduced in recent years, but most of them are based on radio frequency, radar, and other media. We assume that the field of computer vision is mature enough to detect and track invading UAVs. Thus we propose a visible light mode dataset called Dalian University of Technology Anti-UAV dataset, DUT Anti-UAV for short. It contains a detection dataset with a total of 10,000 images and a tracking dataset with 20 videos that include short-term and long-term sequences. All frames and images are manually annotated precisely. We use this dataset to train several existing detection algorithms and evaluate the algorithms’ performance. Several tracking methods are also tested on our tracking dataset. Furthermore, we propose a clear and simple tracking algorithm combined with detection that inherits the detector’s high precision. Extensive experiments show that the tracking performance is improved considerably after fusing detection, thus providing a new attempt at UAV tracking using our dataset. The datasets and results are publicly available at: <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/wangdongdut/DUT-Anti-UAV</uri> .

Research on UAV target detection and substation equipment status recognition technology based on computer vision

The use of computer vision-related technologies can more efficiently achieve target detection of UAVs and accurately identify the status of substation equipment. Therefore, this paper proposes research on UAV target detection and substation equipment status recognition technology based on computer vision. Taking key technical links such as image preprocessing based on computer vision technology, image feature extraction, and construction of UAV target detection coordinate system as breakthrough points, realizing UAV target detection work lays a solid foundation for the identification of substation equipment status. It lays the foundation and promotes the scientific and effective identification of the status of substation equipment. It provides a guarantee for improving the accuracy of UAV target detection and promoting the efficient and stable identification of substation equipment status. CLC number: TN919.8 Document identification code: A.

Manipal-UAV person detection dataset: A step towards benchmarking dataset and algorithms for small object detection

Intelligent UAV video analysis has drawn the attention of many researchers due to the increasing demand for unmanned aerial vehicles (UAVs) in computer vision-related applications. Applications such as search and rescue, the military, and surveillance demand automatic detection of human targets in large-scale UAV images, which is very challenging due to the small size and inadequate feature representation of person objects. Despite the significant advancements in generic object detection tasks, the performance of the state-of-the-art small object detection algorithms falls below the satisfactory level due to the lack of a representative dataset and the limited information available for small objects. To facilitate advancements in UAV and small object detection research, we present a Manipal-UAV person detection dataset11https://github.com/Akshathakrbhat/Manipal-UAV-Person-Dataset. collected from two UAVs flying at varying altitudes, locations, and weather conditions. The dataset contains 13,462 sampled images from 33 videos having 1,53,112 person object instances. The videos are captured in an unconstrained environment with complex scenes covering small objects of varying scales, poses, illumination, and occlusion, making person detection extremely challenging on this newly created dataset. This article compares the characteristics of the Manipal-UAV dataset with the standard VisDrone and Okutama datasets having aerial view person objects. In addition, it provides baseline evaluation results of the various state-of-the-art object detection algorithms applied to the newly created Manipal-UAV Person detection dataset. The dataset is made publicly available at https://github.com/Akshathakrbhat/Manipal-UAV-Person-Dataset.

A Probabilistic–Geometric Approach for UAV Detection and Avoidance Systems

This paper proposes a collision avoidance algorithm for the detection and avoidance capabilities of Unmanned Aerial Vehicles (UAVs). The proposed algorithm aims to ensure minimum separation between UAVs and geofencing with multiple no-fly zones, considering the sensor uncertainties. The main idea is to compute the collision probability and to initiate collision avoidance manoeuvres determined by the differential geometry concept. The proposed algorithm is validated by both theoretical and numerical analysis. The results indicate that the proposed algorithm ensures minimum separation, efficiency, and scalability compared with other benchmark algorithms.

UAV Target Detection under Complex Sky Background

&lt;p class="15"&gt;At present, unmanned aerial vehicles (UAVs) are widely used in various fields, and the management of UAVs is very important to solve the problems in the field of low-altitude safety. Due to the low flying height, small radar cross section, and inconspicuous characteristic signals of UAVs, the detection of UAVs based on video frames taken by fixed cameras cannot meet the existing requirements in terms of tracking speed and recognition accuracy. This paper proposes a multi-sensor fusion model. Firstly, the UAV target signal is improved by spatial filtering and improved Sobel operator edge detection algorithm, and then Gaussian filter is used to denoise, and finally the UAV small target is extracted based on the maximum inter-class variance method threshold segmentation algorithm. Experimental results show that this method can effectively enhance the UAV target signal in a complex environment, and the threshold segmentation method also has good adaptability, and can effectively screen UAVs under a complex sky background.&lt;/p&gt;

Acoustic UAV detection method based on blind source separation framework

Many new challenges are encountered in public security because of the rapid development of an unmanned aerial vehicle (UAV) technology. In order to meet these challenges brought by UAV, it is necessary to realize timely and accurate UAV detection. In this paper, an acoustic UAV detection method based on blind source separation (BSS) framework is introduced to solve the UAV sound detection problem with multi-source interference. In the framework proposed in this paper, source number estimation is used to determine the type of BSS problem, and three methods are applied to separate the UAV sound from the mixing signal in different BSS situations. In the experiment, a variety of machine learning algorithms are used to detect UAVs. The robustness of the proposed method is tested on different UAVs and different sound features. Experimental results show that the UAV sound detection method based on the BSS framework realizes the effective detection of UAVs in different tests and achieves excellent robustness. Compared with mixed signal and filtered signal, the detection rate of our scheme is more than 90%, whether in the overdetermined, positive-definite, and underdetermined cases.