Recognition of pedestrian trajectories and attributes with computer vision and deep learning techniques DOI
Peter Kok-Yiu Wong, Han Luo, Mingzhu Wang

et al.

Advanced Engineering Informatics, Journal Year: 2021, Volume and Issue: 49, P. 101356 - 101356

Published: July 13, 2021

Language: Английский

Roles of artificial intelligence in construction engineering and management: A critical review and future trends DOI
Yue Pan, Limao Zhang

Automation in Construction, Journal Year: 2020, Volume and Issue: 122, P. 103517 - 103517

Published: Dec. 18, 2020

Language: Английский

Citations

777
Trajectory control of electro-hydraulic position servo system using improved PSO-PID controller DOI
Hao Feng, Wei Ma, Chenbo Yin

et al.

Automation in Construction, Journal Year: 2021, Volume and Issue: 127, P. 103722 - 103722

Published: April 18, 2021

Language: Английский

Citations

123
SODA: A large-scale open site object detection dataset for deep learning in construction DOI

Rui Duan,

Hui Deng, Mao Tian

et al.

Automation in Construction, Journal Year: 2022, Volume and Issue: 142, P. 104499 - 104499

Published: July 31, 2022

Language: Английский

Citations

103
Challenges, tasks, and opportunities in teleoperation of excavator toward human-in-the-loop construction automation DOI Creative Commons
Jin Sol Lee, Youngjib Ham, Hangue Park

et al.

Automation in Construction, Journal Year: 2022, Volume and Issue: 135, P. 104119 - 104119

Published: Jan. 10, 2022

Language: Английский

Citations

85
A review of machine learning-based human activity recognition for diverse applications DOI
Farzana Kulsoom, Sanam Narejo, Zahid Mehmood

et al.

Neural Computing and Applications, Journal Year: 2022, Volume and Issue: 34(21), P. 18289 - 18324

Published: Aug. 7, 2022

Language: Английский

Citations

76
Development of an Image Data Set of Construction Machines for Deep Learning Object Detection DOI
Bo Xiao, Shih-Chung Kang

Journal of Computing in Civil Engineering, Journal Year: 2020, Volume and Issue: 35(2)

Published: Dec. 3, 2020

Deep learning object detection algorithms have proven their capacity to identify a variety of objects from images and videos in near real-time speed. The construction industry can potentially benefit this machine intelligence by linking with automatically analyze productivity monitor activities safety perspective. However, an effective image data set machines for training deep is not currently available due the limited accessibility images, time-and-labor-intensiveness manual annotations, knowledge base required terms both learning. This research presents case study on developing specifically named Alberta Construction Image Data Set (ACID). In ACID, 10,000 belonging 10 types are manually collected annotated corresponding positions images. To validate feasibility we train using four existing algorithms, including YOLO-v3, Inception-SSD, R-FCN-ResNet101, Faster-RCNN-ResNet101. mean average precision (mAP) 83.0% 87.8% 88.8% 89.2% speed 16.7 frames per second (fps), which satisfies needs most studies field automation construction.

Language: Английский

Citations

127
Vision-Based Method Integrating Deep Learning Detection for Tracking Multiple Construction Machines DOI
Bo Xiao, Shih-Chung Kang

Journal of Computing in Civil Engineering, Journal Year: 2020, Volume and Issue: 35(2)

Published: Dec. 28, 2020

Tracking construction machines in videos is a fundamental step the automated surveillance of productivity, safety, and project progress. However, existing vision-based tracking methods are not able to achieve high precision, robustness, practical processing speed simultaneously. Occlusions illumination variations on sites also prevent from obtaining optimal performance. To address these challenges, this research proposes method, called machine tracker (CMT), track multiple videos. CMT consists three main modules: detection, association, assignment. The detection module detects using deep learning algorithm YOLOv3 each frame. Then association relates results two consecutive frames, assignment produces results. In testing, achieved 93.2% object accuracy (MOTA) 86.5% precision (MOTP) with 20.8 frames per second when tested four proposed was integrated into framework analyzing excavator productivity earthmoving cycles 96.9% accuracy.

Language: Английский

Citations

85
Digital twin: Stability analysis for tower crane hoisting safety with a scale model DOI
Weiguang Jiang, Lieyun Ding, Cheng Zhou

et al.

Automation in Construction, Journal Year: 2022, Volume and Issue: 138, P. 104257 - 104257

Published: April 15, 2022

Language: Английский

Citations

55
Sensing, perception, decision, planning and action of autonomous excavators DOI
Oybek Eraliev,

Kwang-Hee Lee,

Dae-Young Shin

et al.

Automation in Construction, Journal Year: 2022, Volume and Issue: 141, P. 104428 - 104428

Published: June 24, 2022

Language: Английский

Citations

51
Automatic vision-based calculation of excavator earthmoving productivity using zero-shot learning activity recognition DOI
Chen Chen, Bo Xiao, Yuxuan Zhang

et al.

Automation in Construction, Journal Year: 2022, Volume and Issue: 146, P. 104702 - 104702

Published: Dec. 7, 2022

Language: Английский

Citations

49