TY  - CONF
TI  - Time-lapse Image Processing for Estimating Construction Equipment Emissions
C1  - Osaka and Kyoto, Japan
C3  - Proceedings of the 33rd Annual Conference of the International Group for Lean Construction (IGLC 33)
SP  - 1102
EP  - 1113
PY  - 2025
DO  - 10.24928/2025/0168
AU  - Bharathi, Vimal
AU  - Hong, Kepong
AU  - Teizer, Jochen
AD  - Ph.D. Student, Vimal Bharathi, Department of Civil and Mechanical Engineering, Technical University of Denmark, Kongens Lyngby, Denmark, vvebh@dtu.dk, https://orcid.org/0009-0003-9592-0626
AD  - Ph.D. Student, Vimal Bharathi, Department of Civil and Mechanical Engineering, Technical University of Denmark, Kongens Lyngby, Denmark, keho@dtu.dk, https://orcid.org/0009-0005-0018-0579
AD  - Professor, Department of Civil and Mechanical Engineering, Technical University of Denmark, Kongens Lyngby, Denmark, teizerj@dtu.dk, https://orcid.org/0000-0001-8071-895X
ED  - Seppänen, Olli 
ED  - Koskela, Lauri 
ED  - Murata , Koichi 
AB  - Non-road mobile machinery (NRMM) contributes significantly to environmental pollution.  As part of lean construction methodologies, construction equipment emissions can be monitored remotely by analyzing its working states over time. This offers a viable alternative to expensive, simplified, or portable emission monitoring systems. However, the complex and dynamic aspects of construction sites make it difficult for humans to follow equipment activity. This leads to manual and time-consuming site resource management. This paper proposes automating the task of equipment emission monitoring as part of the more significant goals in lean construction by applying Artificial Intelligence (AI) to photography from a time-lapse camera that oversees a construction site. A Deep Learning model is trained to detect several pieces of equipment on the construction site. The proposed method further tracks the resources to predict whether it is operating between consecutive timestamps. Observing the activity state of the equipment, the respective emission values of the equipment are calculated according to historical and calibrated benchmarks. The results demonstrate a promising method applying existing surveillance technology that, while it also automates parts of construction site operations monitoring of lean schedules, simplifies the process of tracking emissions with respect to construction equipment operations on construction sites. Further work is suggested to acquire higher temporal resolution data and improve data labeling, thereby performing semantic segmentation of equipment for better tracking results for lean construction purposes.
KW  - Computer vision
KW  - Deep Learning
KW  - Object detection and tracking
KW  - Time-lapse camera
KW  - Lean construction equipment operation
KW  - Greenhouse gas emissions.
PB  - 
T2  - Proceedings of the 33rd Annual Conference of the International Group for Lean Construction (IGLC 33)
DA  - 2025/06/02
CY  - Osaka and Kyoto, Japan
L1  - http://iglc.net/Papers/Details/2351/pdf
L2  - http://iglc.net/Papers/Details/2351
N1  - Export Date: 02 June 2025
DB  - IGLC.net
DP  - IGLC
LA  - English
ER  -