TY - GEN
T1 - Vision-based construction activity analysis in long video sequences via hidden markov models
T2 - Construction Research Congress 2018: Safety and Disaster Management, CRC 2018
AU - Roberts, Dominic
AU - Golparvar-Fard, Mani
AU - Niebles, Juan Carlos
AU - Gwak, Junyoung
AU - Bao, Ruxiao
N1 - Funding Information:
The authors would like to acknowledge the financial support of the el Patrimonio Autonomo Fondo Nacional de Financiamiento para la Ciencia, la Tecnologia y la Innovacion, Francisco Jose De Caldas under 20 contract RC No. 0394-2012 with Universidad del Norte, the University of Illinois CEE innovation grant, and NSF CMMI 1446765. The support of many members of the RAAMAC lab including Arsalan Heydarian and Milad Memarzadeh in collecting, annotating, and labeling of videos is also appreciated.
Publisher Copyright:
© 2018 American Society of Civil Engineers (ASCE). All rights reserved.
PY - 2018
Y1 - 2018
N2 - This paper presents a new method for detailed activity analysis of dynamic construction resources in highly varying videos obtained from construction site cameras. Toward this goal, we propose a Hidden Markov Model (HMM) that is able to automatically discover and assign sequences of activities that are most discriminative for an observed construction operation. To do so, the algorithm leverages dense trajectory features from a detected dynamic resource (e.g., excavator) in a video. Using these dense trajectory features, we train a Gaussian mixture model (GMM) to estimate the probability density function of each activity with multiple one-versus-all support vector machine classifiers. The proposed HMM also models duration of each activity, and the transition between activities (e.g., "swing bucket loaded" after "load bucket" for earth-moving activities of an excavator). As a proof-of-concept, we train and test our HMM+GMM model on an unprecedented dataset of 10 real-world long video sequences of interacting pairs of excavators and dumptrucks. Our preliminary experimental results on long-sequence activity recognition in presence of noise, occlusions, and scene clutter demonstrate the effectiveness of our method.
AB - This paper presents a new method for detailed activity analysis of dynamic construction resources in highly varying videos obtained from construction site cameras. Toward this goal, we propose a Hidden Markov Model (HMM) that is able to automatically discover and assign sequences of activities that are most discriminative for an observed construction operation. To do so, the algorithm leverages dense trajectory features from a detected dynamic resource (e.g., excavator) in a video. Using these dense trajectory features, we train a Gaussian mixture model (GMM) to estimate the probability density function of each activity with multiple one-versus-all support vector machine classifiers. The proposed HMM also models duration of each activity, and the transition between activities (e.g., "swing bucket loaded" after "load bucket" for earth-moving activities of an excavator). As a proof-of-concept, we train and test our HMM+GMM model on an unprecedented dataset of 10 real-world long video sequences of interacting pairs of excavators and dumptrucks. Our preliminary experimental results on long-sequence activity recognition in presence of noise, occlusions, and scene clutter demonstrate the effectiveness of our method.
UR - http://www.scopus.com/inward/record.url?scp=85048607856&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85048607856&partnerID=8YFLogxK
U2 - 10.1061/9780784481288.017
DO - 10.1061/9780784481288.017
M3 - Conference contribution
AN - SCOPUS:85048607856
T3 - Construction Research Congress 2018: Safety and Disaster Management - Selected Papers from the Construction Research Congress 2018
SP - 164
EP - 173
BT - Construction Research Congress 2018
A2 - Harper, Christofer
A2 - Lee, Yongcheol
A2 - Harris, Rebecca
A2 - Berryman, Charles
A2 - Wang, Chao
PB - American Society of Civil Engineers
Y2 - 2 April 2018 through 4 April 2018
ER -