Article
Environmental Sciences
Huanhuan Wang, Lisheng Jin, Yang He, Zhen Huo, Guangqi Wang, Xinyu Sun
Summary: This paper proposes a detector-tracker integration framework for autonomous vehicle pedestrian tracking. A pedestrian objects detector based on the improved YOLOv7 network is established, and a novel appearance feature extraction network is proposed, which integrates the convolutional structural re-parameterization idea to construct an optimized DeepSORT tracker. Experimental results on MOT17 and MOT20 public datasets show that the framework has high tracking accuracy and outperforms other multi-object tracking algorithms.
Article
Chemistry, Analytical
Tianle Wu, Suyang Zhong, Hao Chen, Xia Geng
Summary: This study proposes a wheat-ear counting method based on a UAV video multi-objective tracking method, which improves counting efficiency through model optimization and algorithm improvement. The results show that the improved model achieves an average precision of 96.2% and a multi-objective tracking accuracy of 75.4%. Verification results demonstrate that this method effectively counts wheat ears.
Article
Environmental Sciences
Meng Luo, Yanan Tian, Shengwei Zhang, Lei Huang, Huiqiang Wang, Zhiqiang Liu, Lin Yang
Summary: This study proposed an improved Faster R-CNN algorithm for tree detection in mining areas, which achieved higher detection accuracy through strategies such as data augmentation, feature pyramid network, and loss function optimization, outperforming popular target detection algorithms.
Article
Environmental Sciences
Qingqing Hong, Ling Jiang, Zhenghua Zhang, Shu Ji, Chen Gu, Wei Mao, Wenxi Li, Tao Liu, Bin Li, Changwei Tan
Summary: This paper proposes a lightweight model for wheat ear FHB detection based on UAV-enabled edge computing. The model utilizes deep learning architectures and edge devices to achieve accurate and real-time detection. By using a lighter network, optimizing the loss function, and enhancing the model's generalization ability, the proposed model performs well in terms of detection accuracy.
Article
Chemistry, Analytical
Diego Benjumea, Alfonso Alcantara, Agustin Ramos, Arturo Torres-Gonzalez, Pedro Sanchez-Cuevas, Jesus Capitan, Guillermo Heredia, Anibal Ollero
Summary: This paper presents a UAV localization system designed for infrastructure inspection, combining multiple stereo cameras with a robotic total station to provide full-state estimation. The system can align and fuse all sensor measurements in real-time, meeting the challenging flight requirements of UAVs in infrastructure inspection scenarios.
Article
Environmental Sciences
Yue Xi, Wenjing Jia, Qiguang Miao, Xiangzeng Liu, Xiaochen Fan, Hanhui Li
Summary: In this paper, we propose a Fine-grained Target Focusing Network (FiFoNet) that effectively selects multi-scale features, blocks background interference, and enhances the representation of small objects. Furthermore, a Global-Local Context Collector (GLCC) is introduced to extract global and local contextual information for improving low-quality representations. Experimental results demonstrate the superior performance of FiFoNet in object detection for UAV images.
Article
Environmental Sciences
Chuanyun Wang, Yang Su, Jingjing Wang, Tian Wang, Qian Gao
Summary: With the development of unmanned aerial vehicle (UAV) technology and swarm intelligence technology, the study focuses on the threat and challenge brought by UAV swarms to low-altitude airspace defense. A dataset named UAVSwarm is manually annotated for UAV swarm detection and tracking, which includes various scenes and types of UAVs. Advanced detection and multi-object tracking models are used for comprehensive testing and performance verification. The experimental results show the dataset's availability and usability for training and testing various UAV detection and swarm tracking tasks.
Article
Chemistry, Analytical
Kwai-Wa Tse, Rendong Pi, Yuxiang Sun, Chih-Yung Wen, Yurong Feng
Summary: Traditional methods for crack inspection in large infrastructures are time-consuming and costly as they require multiple devices and instruments. In this study, we propose a real-time crack inspection system based on unmanned aerial vehicles, which successfully detects and classifies various types of cracks. The system accurately identifies the crack positions in the world coordinate system. Our detector, an improved YOLOv4 with an attention module, achieves 90.02% mean average precision (mAP) and outperforms the YOLOv4-original by 5.23% in terms of mAP. The proposed system is low-cost, lightweight, and not restricted by navigation trajectories. Experimental results demonstrate its robustness and effectiveness in real-world crack inspection tasks.
Article
Computer Science, Information Systems
Naeem Ayoub, Peter Schneider-Kamp
Summary: The use of deep learning-based autonomous drone vision systems shows promising results in detecting faults in power line components, providing an effective solution for real-time on-board power line inspection. Various single-board devices were utilized for experimental evaluation in running deep learning models.
Article
Chemistry, Multidisciplinary
Jin Hong, Junseok Kwon
Summary: This paper proposes a novel visual tracking method for unmanned aerial vehicles in aerial scenery, utilizing a new object proposal method that is robust to small objects and severe background clutter. Experimental results demonstrate accurate tracking of UAVs even in challenging conditions, outperforming conventional methods in both qualitative and quantitative assessments.
APPLIED SCIENCES-BASEL
(2021)
Review
Engineering, Electrical & Electronic
Yuqi Han, Huaping Liu, Yufeng Wang, Chunlei Liu
Summary: Unmanned aerial vehicles have been widely used in military and civilian fields due to their flexibility and efficiency. The vision system, as an essential component of UAVs, has gained significant attention in recent years for various applications. This review focuses on the automatic understanding of visual data collected from UAVs and provides an overview of techniques and developments in object detection, tracking, and semantic segmentation. The challenges and future directions in UAV vision are also highlighted.
IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND REMOTE SENSING
(2022)
Article
Engineering, Electrical & Electronic
Ang Li, Shouxiang Ni, Yanan Chen, Jianxin Chen, Xin Wei, Liang Zhou, Mohsen Guizani
Summary: This paper proposes a cross-modal knowledge distillation (CKD) enabled object detection paradigm for UAV-based target detection. It achieves comparable detection performance with multi-modal techniques while requiring less computational resources.
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
(2023)
Review
Agronomy
Daniel Olson, James Anderson
Summary: The advancement of UAV technologies in agriculture offers producers the ability to enhance efficiency and quality of crop production, as well as the capability to quickly identify and address problematic areas.
Article
Construction & Building Technology
Zheng-fang Wang, Yan-fei Yu, Jing Wang, Jian-qing Zhang, Hong-liang Zhu, Peng Li, Lei Xu, Hao-nan Jiang, Qing-mei Sui, Lei Jia, Jiang-ping Chen
Summary: In this study, a novel convolutional neural network is proposed to automatically identify dam-surface seepage from low-resolution thermograms. The method achieves superior results by reducing false alarms caused by background interference and accurately identifying seepage profiles. Experimental results confirm the effectiveness of the proposed network.
CONSTRUCTION AND BUILDING MATERIALS
(2022)
Article
Computer Science, Information Systems
Lunyuan Chen, Rui Zhao, Ke He, Zichao Zhao, Liseng Fan
Summary: This paper investigates intelligent ubiquitous computing for future UAV-enabled MEC systems, optimizing system performance using reinforcement learning and transfer learning algorithms to reduce latency and energy consumption. Reinforcement learning is used to devise offloading strategy meeting constraints and alleviate effects of jamming attack, while transfer learning speeds up training process and enhances reinforcement learning performance. Simulation results show proposed offloading strategy outperforms conventional methods, and using transfer learning achieves better system performance with significantly reduced training time.
CLUSTER COMPUTING-THE JOURNAL OF NETWORKS SOFTWARE TOOLS AND APPLICATIONS
(2022)
Article
Chemistry, Analytical
Anas M. Ali, Bilel Benjdira, Anis Koubaa, Walid El-Shafai, Zahid Khan, Wadii Boulila
Summary: This study extensively investigates the efficiency of Vision Transformer (ViT) in image restoration. ViT architectures are classified for each task, and the outcomes, advantages, limitations, and possible areas for future research are detailed. Overall, incorporating ViT in new architectures for image restoration is becoming a rule due to its better efficiency, robustness in feature extraction, and improved feature learning approach. However, there are drawbacks such as the need for more data to demonstrate the benefits over CNN, increased computational cost, more challenging training process, and lack of interpretability.
Article
Chemistry, Analytical
Adel Ammar, Anis Koubaa, Wadii Boulila, Bilel Benjdira, Yasser Alhabashi
Summary: This paper presents a novel real-time vehicle identification and license plate recognition system based on video streaming. The system integrates multiple algorithms, including object detectors, image classifiers, and a multi-object tracker, to recognize car models and license plates. By leveraging the information redundancy of Saudi license plates' characters, the system achieves high accuracy while ensuring real-time performance. Experimental results show that the system can achieve a frame rate of 17.1 FPS on a Jetson Xavier AGX edge device in real-world environments.
Article
Environmental Sciences
Anas M. Ali, Bilel Benjdira, Anis Koubaa, Wadii Boulila, Walid El-Shafai
Summary: In this study, a new architecture called TESR was designed to enhance the resolution of remote sensing images using Vision Transformers (ViT) and Diffusion Model (DM). TESR outperformed state-of-the-art techniques on the UCMerced benchmark dataset, improving the image quality in terms of PSNR/SSIM metrics.
Article
Green & Sustainable Science & Technology
Vishnu Kumar Kaliappan, Sundharamurthy Gnanamurthy, Abid Yahya, Ravi Samikannu, Muhammad Babar, Basit Qureshi, Anis Koubaa
Summary: Smart healthcare utilizes cloud computing and the Internet of Things to achieve remote patient monitoring, real-time data collection, enhanced data security, and cost-effective storage and analysis of healthcare data. This paper presents an information-centric dissemination scheme (ICDS) designed for smart healthcare services in smart cities. The proposed scheme addresses the time sensitivity of healthcare data and aims to ensure consistent dissemination. The ICDS utilizes decision-tree learning to classify requests based on time-sensitive features, enabling access prioritization. The scheme also involves segregating sensitive information and distributing digital health data within the classified time to maintain time sensitivity and prioritize access. The learning is then adapted for the leaves based on data significance and minimum resources to reduce waiting times and enhance availability.
Article
Chemistry, Analytical
Nesrine Atitallah, Omar Cheikhrouhou, Khaleel Mershad, Anis Koubaa, Fahima Hajjej
Summary: This paper introduces a power-conscious routing approach for resource-restricted wireless sensor networks (WSNs), leveraging cooperative communications and innovative relay node selection techniques. The proposed cooperative and efficient routing protocol (CERP) enhances reliability and energy conservation, outperforming the standard RPL protocol by 10% in reliability and 18% in energy savings.
Article
Mathematics, Interdisciplinary Applications
Farman Ali Shah, Wadii Boulila, Anis Koubaa, Nabil Mlaiki
Summary: This study presents a highly accurate numerical method for solving the advection-diffusion equation of fractional order. The proposed method uses the Laplace transform to handle the time-fractional derivative and utilizes the Chebyshev spectral collocation method for spatial discretization. The method shows high accuracy and stability, as demonstrated by solving various problems in two dimensions.
FRACTAL AND FRACTIONAL
(2023)
Article
Computer Science, Information Systems
Dennies Tsietso, Abid Yahya, Ravi Samikannu, Muhammad Usman Tariq, Muhammad Babar, Basit Qureshi, Anis Koubaa
Summary: Breast cancer is a leading cause of death among women worldwide, and early detection is crucial for reducing mortality. This paper proposes a novel computer-aided diagnosis (CADx) system that uses deep learning techniques and incorporates multiple breast thermogram views and patient clinical data to improve accuracy. The system outperforms single-input models and achieves an overall accuracy of 90.48%, a sensitivity of 93.33%, and an AUROC curve of 0.94. This approach could provide a more cost-effective and less hazardous screening option for breast cancer detection, particularly for diverse age groups.
Article
Computer Science, Information Systems
Safi Ullah, Wadii Boulila, Anis Koubaa, Jawad Ahmad
Summary: The increasing prevalence of the Industrial Internet of Things (IIoT) in industrial environments amplifies the potential for security breaches. To address challenges in training intrusion detection systems (IDS) for IIoT networks, this paper proposes a method called multi-head attention-based gated recurrent unit (MAGRU) that utilizes machine learning and deep learning to detect malicious activities. The proposed MAGRU outperforms other models, achieving high precision, recall, F1-score, and accuracy in detecting intrusions in IIoT networks.
Proceedings Paper
Computer Science, Artificial Intelligence
Adeel Zaidi, Muhammad Kazim, Lixian Zhang, Ahmad Taher Azar, Anis Koubaa, Bilel Benjdira, Adel Ammar, Mohammad Abdelkader
Summary: This research paper proposes a combination of a deep neural network algorithm and a cascaded PID+FF controller for safe and obstacle-free aerial transportation. The results show that this approach improves accuracy and ensures flight safety in highly dynamic environments.
2022 2ND INTERNATIONAL CONFERENCE OF SMART SYSTEMS AND EMERGING TECHNOLOGIES (SMARTTECH 2022)
(2022)
Article
Computer Science, Information Systems
Yasir Shahzad, Huma Javed, Haleem Farman, Zahid Khan, Moustafa M. Nasralla, Anis Koubaa
Summary: This study proposes an optimal distributive cross-layer and thermal-aware convergecast protocol to address the limitations of current protocols in terms of convergence delays, communication and processing overheads. The proposed protocol accelerates the convergence process, improves energy efficiency and thermal control, and supports a distributive hierarchy.