Article
Environmental Sciences
Xianjiang Li, Boyong He, Kaiwen Ding, Weijie Guo, Bo Huang, Liaoni Wu
Summary: This paper presents a wide-area and real-time object search system based on deep learning for unmanned aerial vehicles (UAVs). By utilizing parallel systems, simplified algorithms, and optimization techniques like TensorRT, the system achieves efficient and real-time object detection, as demonstrated through practical application.
Article
Geochemistry & Geophysics
Yong Zhao, Lin Chen, Xishan Zhang, Shibiao Xu, Shuhui Bu, Hongkai Jiang, Pengcheng Han, Ke Li, Gang Wan
Summary: Inspired by SLAM workflow, the article presents an online sequential SfM solution for high-frequency video and high-resolution aerial images, achieving high efficiency and precision. The method optimizes camera poses, generates dense point clouds in real time, and demonstrates robust performance in aerial image mosaic and DSM reconstruction in real time.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Environmental Sciences
Seungmin Lee, Dat Ngo, Bongsoon Kang
Summary: This paper highlights the significance of image dehazing technology in photography, computer vision, and remote sensing applications. It presents a non-deep learning autonomous image dehazing algorithm and a corresponding FPGA design for high-quality real-time vision systems. Extensive experiments are conducted to validate the efficacy of the proposed design across different aspects, and a method for synthesizing cloudy images is introduced to facilitate future aerial surveillance research.
Article
Engineering, Aerospace
Fang Wang, Xiaoyan Luo, Qixiong Wang, Lu Li
Summary: A lightweight and dual-path deep convolutional architecture, Aerial-BiSeNet, is proposed for real-time segmentation on high-resolution aerial images, achieving leading performance in terms of both accuracy and efficiency.
CHINESE JOURNAL OF AERONAUTICS
(2021)
Article
Chemistry, Analytical
Bernardo Santana, El Khalil Cherif, Alexandre Bernardino, Ricardo Ribeiro
Summary: The application of aerial vehicle images for observing fires in forest environments is limited by the lack of distinctive visual features. To address this issue, a real-time forest fire georeferencing and filtering algorithm was developed, which showed promising results in terms of georeferencing accuracy and filtering performance.
Article
Construction & Building Technology
Min-Lung Cheng, Masashi Matsuoka, Wen Liu, Fumio Yamazaki
Summary: This paper introduces a systematic workflow for on-the-fly 3D reconstruction in disaster areas using optical imagery acquired by drones. It proposes a strategy for spatially linking sequential images and determining suitable stereopair selection. The study also develops criteria for valid epipolar stereoapair determination to improve the effectiveness of 3D dense reconstruction. The research utilizes a dataset from collapsed buildings induced by the 2016 Kumamoto earthquake in Japan to simulate more effective 3D reconstruction, achieving a mean data processing time of approximately ten seconds per image.
AUTOMATION IN CONSTRUCTION
(2022)
Article
Computer Science, Information Systems
Honglian Wang, Peiyan Li, Yang Liu, Junming Shao
Summary: This paper proposes a new method for next point-of-interest (POI) recommendation, called DSPR, by exploring user preferences and real-time demand simultaneously to support the final POI recommendation. Experimental results show that DSPR outperforms many state-of-the-art methods in recommendation performance.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Information Systems
Danilo Avola, Luigi Cinque, Alessio Fagioli, Gian Luca Foresti, Daniele Pannone, Claudio Piciarelli
Summary: The study proposes a method to automatically estimate UAV monitoring parameters, successfully applied to tasks including area mosaicking, change detection over time, and ground content classification. Parameter estimation involves factors such as target size, UAV cruise speed, and internal parameters of the video sensor.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Geography, Physical
Jae-In Kim, Chang-Uk Hyun, Hyangsun Han, Hyun-Cheol Kim
Summary: This paper proposes a robust method to generate high-quality DSMs for drifting sea ice, introducing improvements like relative georeferencing, match inspection, adaptive search-window adjustment, and robust vertical positioning. Experimental results show significant quality enhancements compared with existing methods.
ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING
(2021)
Article
Optics
Li Ruixia, Hu Shanting, Gu Xiaodong, Fumio Koyama
Summary: In this paper, a real-time scanning structured-light depth sensing system based on a solid-state VCSEL beam scanner and an electro-thermally tunable VCSEL is demonstrated. The system is capable of capturing real-time depth images with a wide field of view and high resolution. By using a higher-speed camera, even higher resolution and faster frame rates can be achieved.
Article
Computer Science, Hardware & Architecture
Alexander Menshchikov, Dmitrii Shadrin, Viktor Prutyanov, Daniil Lopatkin, Sergey Sosnin, Evgeny Tsykunov, Evgeny Iakovlev, Andrey Somov
Summary: The Hogweed of Sosnowskyi, a fast-growing plant with toxic properties, has spread widely across Eurasia, from Germany to Siberia. Utilizing an Unmanned Aerial Vehicle (UAV) equipped with Fully Convolutional Neural Networks (FCNN), a model has been developed to achieve accurate detection and segmentation of hogweed, allowing for comprehensive data collection to combat its expansion.
IEEE TRANSACTIONS ON COMPUTERS
(2021)
Article
Geochemistry & Geophysics
Haris Ijaz, Rizwan Ahmad, Rehan Ahmed, Waqas Ahmed, Yan Kai, Wu Jun
Summary: Unmanned aerial vehicles equipped with onboard embedded platforms and camera sensors offer crucial autonomous decision-making capabilities in disaster recovery and management. To achieve real-time disaster scenario classification, a framework using UAVs for edge computation is proposed, and throughput is increased through optimized model compression.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2023)
Article
Engineering, Electrical & Electronic
Junjie Lu, Bailing Tian, Hongming Shen, Xuewei Zhang
Summary: In this article, a real-time instance-aware segmentation and semantic mapping method for UAVs on small edge devices is proposed. By using a lightweight object detection model as the backbone and reformulating the mask generation problem as threshold regression in depth by a novel designed truncation network, the proposed method achieves a speed of 38 frames/s. Autonomous exploration experiments of UAVs demonstrate the effectiveness of the method in both simulation and real-world.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Article
Energy & Fuels
Chuanyang Liu, Yiquan Wu, Jingjing Liu, Jiaming Han
Summary: The study proposes the MTI-YOLO network for insulator detection in complex aerial images, achieving improved accuracy through techniques such as multi-scale feature detection and spatial pyramid pooling. Experimental results show that the proposed network outperforms traditional methods in both complex backgrounds and bright illumination conditions.
Article
Computer Science, Artificial Intelligence
Jipeng Wu, Rongrong Ji, Jianzhuang Liu, Mingliang Xu, Jiawen Zheng, Ling Shao, Qi Tian
Summary: This paper introduces a novel Sequential Prediction Network (SPNet) with Spatial Semantic and Edge Loss (SEL) and an adversarial network to achieve high segmentation accuracy in real-time applications. By utilizing a knowledge distillation scheme, the method effectively compresses structured knowledge from cumbersome networks, achieving promising results.