Air-to-Air Visual Detection of Micro-UAVs: An Experimental Evaluation of Deep Learning

Zheng, Ye; Chen, Zhang; Lv, Dailin; Li, Zhixing; Lan, Zhenzhong; Zhao, Shiyu

doi:10.1109/lra.2021.3056059

Cited by 88 publications

(27 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, using PTZ cameras is not suitable in congested environment where a LOS to the target drone is often lost. An alternate approach is the airto-air visual detection [121], which uses a drone detection and tracking using another flying drone equipped with a camera and onboard detection system. The existing datasets that typically contain drones images captured from a ground camera may not provide sufficient accuracy due to the fact that drones' appear different when seen from a flying drone.…”

Section: Visual Detectionmentioning

confidence: 99%

“…The existing datasets that typically contain drones images captured from a ground camera may not provide sufficient accuracy due to the fact that drones' appear different when seen from a flying drone. Hence, authors in [121] developed a large dataset i.e., "Det-Fly" containing 13,000 images of drones captured from another flying drones. 3) Summary and Insights: Visual detection despite its inherited limitations (e.g., LOS and proper lighting requirements) is gaining popularity in the academic research likely due of the success and growing interest of researchers in the area of DL and computer vision.…”

Section: Visual Detectionmentioning

confidence: 99%

See 1 more Smart Citation

On the Detection of Unauthorized Drones—Techniques and Future Perspectives: A Review

et al. 2022

View full text Add to dashboard Cite

The market size of civilian drones is tremendously increasing and is expected to reach 1.66 million by the end of 2023. The increase in number of civilian drones poses several privacy and security threats. To safeguard critical assets and infrastructure and to protect privacy of people from the illegitimate uses of commercial drones, a drone detection system is inevitable. In particular, there is a need for a drone detection system that is efficient, accurate, robust, costeffective and scalable. Recognizing the importance of the problem, several drone detection approaches have been proposed over time. However, none of these provides sufficient performance due to the inherited limitations of the underlying detection technology. More specifically, there are trade-offs among various performance metrics e.g., accuracy, detection range, and robustness against environmental conditions etc. This motivates an in-depth study and critical analysis of the existing approaches, highlighting their potential benefits and limitations. In this paper, we provide a rigorous overview of the existing drone detection techniques and a critical review of the state-of-the-art. Based on the review, we provide key insights on the future drone detection systems. We believe these insights will provide researchers and practicing engineers a holistic view to understand the broader context of the drone detection problem.

show abstract

Section: Visual Detectionmentioning

confidence: 99%

Section: Visual Detectionmentioning

confidence: 99%

On the Detection of Unauthorized Drones—Techniques and Future Perspectives: A Review

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Mel-frequency and linear predictive cepstral coefficients with support vector machines (SVM) achieved 96.7% accuracy [2]. Zheng et al in [4] utilized the monocular cameras to perform AAD on micro UAVs, which proved vital for vision-based swarm and malicious UAV detection. They introduced Det-Fly, a novel dataset containing 13,000 images of a flying target UAV in multiple circumstances.…”

Section: B Contributionsmentioning

confidence: 99%

“…As a consequence, the background of the target UAV would have complex scenes such as urban or natural settings. As the onboard camera is flying dynamically, the appearance of the target UAV, such as its form, scale, and color, may change significantly [4], and this method results in less performance compared to GAD.…”

Section: Introductionmentioning

confidence: 99%

“…According to Lykou et al in [5], around 6% commercial UAV detection systems designed on acoustic sensors, 26% are using radio frequency (RF), 28% are radar-based, and 40% adopted visual sensors for detection. Mostly those sensors adopted for drone detection and classification are: radar (radio detection and ranging on several different frequency bands for both active and passive sensing) [6], cameras for the visible spectrum sensing [4], cameras detecting thermal or infrared (IR) emissions [7], LIDAR (light detection and ranging) [8], microphones/acoustic sensors for acoustic vibration detection [2], and RF [9], [10] monitoring sensors for the detection of radio signals [11].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

SafeSpace MFNet: Precise and Efficient MultiFeature Drone Detection Network

dil¹,

khan²,

Alam³

et al. 2022

Preprint

View full text Add to dashboard Cite

<p>Unmanned air vehicles (UAVs) popularity is on the rise as it enables the services like traffic monitoring, emergency communications, deliveries, and surveillance. However, the unauthorized usage of UAVs (a.k.a drone) may violate security and privacy protocols for security-sensitive national and international institutions. The presented challenges require fast, efficient, and precise detection of UAVs irrespective of harsh weather conditions, the presence of different objects, and their size to enable SafeSpace. Recently, there has been significant progress in using the latest deep learning models, but those models have shortcomings in terms of computational complexity, precision, and non-scalability. To overcome these limitations, we propose a precise and efficient multiscale and multifeature UAV detection network for SafeSpace, i.e., \textit{MultiFeatureNet} (\textit{MFNet}), an improved version of the popular object detection algorithm YOLOv5s. In \textit{MFNet}, we perform multiple changes in the backbone and neck of the YOLOv5s network to focus on the various small and ignored features required for accurate and fast UAV detection. To further improve the accuracy and focus on the specific situation and multiscale UAVs, we classify the \textit{MFNet} into small (S), medium (M), and large (L): these are the combinations of various size filters in the convolution and the bottleneckCSP layers, reside in the backbone and neck of the architecture. This classification helps to overcome the computational cost by training the model on a specific feature map rather than all the features. The results show significant performance gain even for unseen feature maps with minimal loss in accuracy. Results show a significant reduction in training parameters, inference, and increased pattern in FPS and GFLOPs for \textit{MFNet} compared to YOLOv5s. \textit{MFNet-M} performance evaluation in terms of precision, recall, mean average-precision (mAP), and IOU increased around 1.8\%, 2.2\%, 0.9\%, 1.7\% compared to YOLOv5s. Furthermore, \textit{MFNet-M} achieves the best performance with 96.8\% precision, 88.4\% recall, 95.9\% mAP, and 51.1\% IoU for UAV detection. The dataset and code are available as an open source: github.com/ZeeshanKaleem/MultiFeatureNet.</p>

show abstract

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

Ding

Guo

Peng

et al. 2023

Advanced Intelligent Systems

View full text Add to dashboard Cite

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.

show abstract

Air-to-Air Visual Detection of Micro-UAVs: An Experimental Evaluation of Deep Learning

Cited by 88 publications

References 31 publications

On the Detection of Unauthorized Drones—Techniques and Future Perspectives: A Review

On the Detection of Unauthorized Drones—Techniques and Future Perspectives: A Review

SafeSpace MFNet: Precise and Efficient MultiFeature Drone Detection Network

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

Contact Info

Product

Resources

About