Article
Environmental Sciences
Christoph Zambanini, Volker Reinprecht, Daniel Scott Kieffer
Summary: InSARTrac is an innovative method that combines InSAR and computer vision-based feature tracking for 3D displacement monitoring. It provides more accurate measurements compared to 1D or 2D data and allows evaluations of kinematics mechanisms. The study presents the results of InSARTrac measurements at Molltal Glacier in Austria, showing a mean displacement rate of 22 mm/day with vertical displacements of 6 to 18 mm/day. The accuracy of InSARTrac is 4.2 ppm, indicating its potential for monitoring various geologic phenomena.
Article
Engineering, Civil
Hao Yang, Jiarui Cai, Meixin Zhu, Chenxi Liu, Yinhai Wang
Summary: The article introduces a novel framework called TIMS system for network-level traffic information estimation, which enhances the accuracy of cross-camera information estimation through the redesign of vision-based vehicle ReID method.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Wenyu Yang, Yong Jiang, Shuai Wen, Yong Fan
Summary: In existing online multiple object tracking algorithms, a scheme that combines object detection and re-identification tasks in a single model has gained attention for its balanced speed and accuracy. However, learning two different tasks in the same model can lead to competition and hinder optimal performance. To address this, a task-related attention network is proposed, along with a smooth gradient-boosting loss function, improving the quality of ReID features. Extensive experiments on MOT16, MOT17, and MOT20 datasets demonstrate the effectiveness of the proposed method, which is also competitive with current mainstream algorithms.
IET COMPUTER VISION
(2023)
Article
Engineering, Multidisciplinary
Xijun Ye, Yongjie Cao, Airong Liu, Xinwei Wang, Yinghao Zhao, Nan Hu
Summary: This study proposes a parallel convolutional neural network (P-CNN) that can extract multidimensional features for damage detection. The P-CNN outperforms traditional CNNs in terms of feature separation and accuracy for structural damage detection. It also exhibits robust performance in high signal-to-noise ratio environments and surpasses other commonly used methods in damage identification.
STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL
(2023)
Review
Chemistry, Analytical
Yizhou Zhuang, Weimin Chen, Tao Jin, Bin Chen, He Zhang, Wen Zhang
Summary: This paper studied the application of computer vision-based structural deformation monitoring techniques in the field of structural health monitoring (SHM) and analyzed the influence mechanism of the measuring accuracy from two perspectives, providing solutions for it.
Article
Chemistry, Multidisciplinary
Xin Duan, Xi Chu, Weizhu Zhu, Zhixiang Zhou, Rui Luo, Junhao Meng
Summary: A method for structural full-field displacement monitoring and damage identification under natural texture conditions is proposed in this study. The feature points of the structure are extracted using the image scale-invariant feature transform, and a calculation theory for the structure's full-field displacement vector is established based on the analysis of the feature points' relative position change. The validation results show that the proposed method exhibits good performance in damage identification.
APPLIED SCIENCES-BASEL
(2023)
Review
Computer Science, Information Systems
Nur Syazarin Natasha Abd Aziz, Salwani Mohd Daud, Rudzidatul Akmam Dziyauddin, Mohamad Zulkefli Adam, Azizul Azizan
Summary: The review highlights the importance of productivity and profitability in poultry farming, emphasizing the role of computer vision technology in optimizing performance and economic management. The paper recommends focusing on key works related to computer vision in poultry farms to address future challenges and ensure successful implementation in the industry.
Article
Engineering, Electrical & Electronic
Laura Ruotsalainen, Aiden Morrison, Maija Makela, Jesperi Rantanen, Nadezda Sokolova
Summary: Collaborative navigation is an infrastructure-free indoor navigation technique that shows promise for a group of pedestrians. Using inertial sensors and a camera with visual odometry, this method addresses the challenges of navigation environment and measurement errors. By employing a depth camera and deep learning based detector, the accuracy of the solution is improved.
IEEE SENSORS JOURNAL
(2022)
Article
Engineering, Civil
Ming Xia, Junjie Lin, Linghao Ying, Jian Sun, Kaikai Chi, Kun Gao, Keping Yu
Summary: The risk of death and injury from traffic crashes is recognized as a serious threat to sustainable development. Aggressive driving, characterized by excessive lane changes, is prevalent and monitoring lane changes can improve transportation sustainability. This article presents BackWatch, a novel vehicle-mounted sensing system that uses a back view cabin camera to monitor steering wheel rotations and track lane-change events. The system achieves high precision and recall in detecting lane changes.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2022)
Article
Construction & Building Technology
Yufeng Zhang, Junxin Xie, Jiayi Peng, Hui Li, Yong Huang
Summary: In this study, a vehicle re-identification method based on Deep Learning and HardNet is proposed for structural health monitoring of bridge structures. Comparative experiments demonstrate the superior performance of this method on vehicle image data.
ADVANCES IN STRUCTURAL ENGINEERING
(2021)
Article
Engineering, Multidisciplinary
Yuequan Bao, Hui Li
Summary: The conventional vibration-based methods face challenges in accurately detecting structural damages, thus necessitating the development of novel diagnosis and prognosis methods based on various monitoring data.
STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL
(2021)
Article
Computer Science, Information Systems
Lang Deng, Jianfei Yang, Shenghai Yuan, Han Zou, Chris Xiaoxuan Lu, Lihua Xie
Summary: In this article, a novel multimodal gait recognition method called GaitFi is proposed, which combines WiFi signals and videos for human identification. The GaitFi outperforms state-of-the-art gait recognition methods and achieves excellent results in real-world experiments.
IEEE INTERNET OF THINGS JOURNAL
(2023)
Article
Computer Science, Artificial Intelligence
Matthieu Bovery, Geraldine Dawson, Jordan Hashemi, Guillermo Sapiro
Summary: This study developed a low-cost alternative method for measuring attention in children with ASD, showing that they exhibit less attention and prefer non-social stimuli over social stimuli. The integration of stimuli design and automatic response analysis using off-the-shelf cameras provides an opportunity to assess behavioral biomarkers.
IEEE TRANSACTIONS ON AFFECTIVE COMPUTING
(2021)
Article
Agriculture, Multidisciplinary
Claire Flemmer, Huub Bakker, Rory Flemmer
Summary: This study introduces a method to precisely measure the roll, pitch, and yaw of tumbling apples and provide detailed statistical descriptions of the tumbling process. The research includes testing on four different apple varieties, presenting probability histograms and fitting Skew-Gaussian distributions to the data. The stochastic characterisations developed in this study can be used for future Monte Carlo simulations to accurately determine camera coverage during automated inspection of apples tumbling on rollers, making a significant contribution to the field.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2021)
Article
Ecology
Kim Bjerge, Hjalte M. R. Mann, Toke Thomas Hoye
Summary: An intelligent camera system is developed to detect, track, and identify individual insects in situ, utilizing computer vision techniques for real-time monitoring and classification. The system shows promising results in non-destructive and real-time monitoring of insects, offering novel insights into the ecology of flower visiting insects.
REMOTE SENSING IN ECOLOGY AND CONSERVATION
(2022)
Article
Construction & Building Technology
Vedhus Hoskere, Jong-Woong Park, Hyungchul Yoon, Billie F. Spencer
JOURNAL OF STRUCTURAL ENGINEERING
(2019)
Article
Chemistry, Analytical
Jongbin Won, Jong-Woong Park, Kyoohong Park, Hyungchul Yoon, Do-Soo Moon
Article
Chemistry, Analytical
Hyungchul Yoon, Kevin Han, Youngjib Ham
Article
Chemistry, Multidisciplinary
Jae Hyuk Lee, Jeong Jun Park, Hyungchul Yoon
APPLIED SCIENCES-BASEL
(2020)
Article
Chemistry, Multidisciplinary
Gun Park, Hyungchul Yoon, Ki-Nam Hong
APPLIED SCIENCES-BASEL
(2020)
Article
Chemistry, Multidisciplinary
Gun Park, Ki-Nam Hong, Hyungchul Yoon
Summary: This paper introduces a new computer-vision based method for structural FE model updating using genetic algorithm. The system measures the displacement of the structure using seven different object tracking algorithms, and optimizes the structural parameters using genetic algorithm. The proposed method successfully updated the FE model to the damaged state and is expected to reduce the time and cost of FE model updating.
APPLIED SCIENCES-BASEL
(2021)
Article
Green & Sustainable Science & Technology
Seongnoh Ahn, Gun Park, Hyungchul Yoon, Jae-Hyeok Han, Jongwon Jung
Summary: Modeling soil-structure interaction in seismic design involves using soil response curves for SDOF structures, but real structures are MDOF. This study compared shaking-table-derived p-y curves for SDOF and MDOF superstructures, finding that increasing DOF had diminishing effects at greater pile depths. Numerical analysis results based on natural periods and mass participation rates of structures were similar to shaking table tests.
Article
Chemistry, Multidisciplinary
Gun Park, Jae Hyuk Lee, Hyungchul Yoon
Summary: Current maintenance practices are time-consuming and costly, necessitating the need for a new maintenance technique. Building information modeling (BIM) has been applied to systematic planning and design in construction, but its application to existing structures is limited. This study proposes a method using photographic data and unmanned aerial vehicles to automatically construct 3D models, potentially improving current maintenance procedures.
APPLIED SCIENCES-BASEL
(2021)
Article
Chemistry, Analytical
Geonhee Lee, Sunjoong Kim, Sangsub Ahn, Ho-Kyung Kim, Hyungchul Yoon
Summary: This study introduces a new vision-based cable displacement measurement system that can measure the displacement of a cable even when the camera is installed in the side of the cable. The proposed method was validated through simulation-based tests, lab-scale tests, and on-site tests.
Article
Construction & Building Technology
Geonyeol Jeon, Sunjoong Kim, Sangsub Ahn, Ho-kyung Kim, Hyungchul Yoon
Summary: Cables are crucial structural components of cable-stayed bridges, but their vibrations caused by wind and vehicles can greatly affect the bridges' usability. To overcome the challenges of installing sensors for measuring cable vibrations, researchers have introduced a computer-vision-based displacement method. However, manual selection of the region of interest (ROI) and loss of feature points remain issues in this method. This paper proposes a new method that automatically selects the ROI using a convolutional neural network and tracks feature points more robustly using a modified Kanade-Lucas-Tomasi (KLT) algorithm.
STRUCTURAL CONTROL & HEALTH MONITORING
(2022)
Article
Construction & Building Technology
Yunwoo Lee, Geonhee Lee, Do Soo Moon, Hyungchul Yoon
Summary: While structural displacements are crucial for structural health monitoring, traditional methods are not widely used due to inconvenience. Recently, vision-based displacement measurement methods have been introduced, offering greater convenience and cost-effectiveness. This study proposes a methodology for measuring structural displacements based on relative view changes induced by camera motion.
STRUCTURAL CONTROL & HEALTH MONITORING
(2022)
Article
Multidisciplinary Sciences
Gun Park, Jongwon Jung, Hyungchul Yoon
Summary: A finite element model updating method considering soil-structure interaction was developed to analyze the effect of soil properties on structural response. LS-DYNA, a commercial finite element program, was included in the loop of the proposed technique using MATLAB. A large-scale shake table test was conducted to validate the performance of the method. The proposed method achieved a maximum accuracy of 91% in estimating structure stiffness, while the conventional method achieved 88%. By comparing the two methods, it was confirmed that the proposed method had an average 3% higher accuracy.
SCIENTIFIC REPORTS
(2023)
Article
Engineering, Civil
Hun Lee, Hyungchul Yoon, Sunjoong Kim
Summary: This study proposes a cost-effective solution to monitor excessive vibrations in stay-cables of long-span bridges using surveillance cameras. Deep learning and computer vision techniques are employed to address technical challenges and the proposed method offers a viable alternative to traditional monitoring methods.
ENGINEERING STRUCTURES
(2023)
Article
Engineering, Civil
Jaehyung Park, Jongwon Jung, Seunghee Park, Hyungchul Yoon
Summary: Vibration-based structural health monitoring ensures structural safety by installing sensors. The peak picking method analyzes the dynamic characteristics of a structure using the peaks of the frequency response function, but results may vary depending on the person predicting the peak point and noise can affect accuracy. To overcome these limitations, this study proposes a new method that uses LSTM networks to automate modal analysis and improves accuracy by considering phase and amplitude information.
SMART STRUCTURES AND SYSTEMS
(2023)
Article
Engineering, Civil
Jeonghyeok Lim, Hyungchul Yoon
Summary: Recently, there has been a global increase in the construction of long span pedestrian suspension bridges. However, the lack of specific construction standards and safety inspections have raised concerns about their safety. This study aims to develop a system identification method for pedestrian suspension bridges that considers both the input and output of the dynamic system. By utilizing artificial intelligence and computer vision techniques, this method estimates the location and magnitude of the pedestrian load, as well as the dynamic response of the bridges. The proposed method is expected to improve the accuracy and efficiency of current inspection and monitoring systems for pedestrian suspension bridges.
SMART STRUCTURES AND SYSTEMS
(2022)
