Action Recognition of Construction Machinery from Simulated Training Data Using Video Filters

Sim, Jinhyeok; Kasahara, Jun Younes Louhi; Chikushi, Shota; Yamakawa, Hiroshi; Tamura, Yusuke; Nagatani, Keiji; Chiba, Takumi; Yamamoto, Shozo; Chayama, Kazuhiro; Yamashita, Atsushi; Asama, Hajime

doi:10.22260/isarc2020/0083

Cited by 3 publications

(2 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Some researchers also used CNN to capture information on-site, which can be used for simulation purposes to further improve construction productivity (Park et al, 2021). Meanwhile, various deep learning techniques are also applied to track workers (Fang et al, 2018a(Fang et al, , 2018b(Fang et al, , 2018cSon et al, 2019) and machinery Shih-Chung, 2021a, 2021b;Dong et al, 2022;Rashid and Louis, 2019;Roberts and Golparvar-Fard, 2019;Sim et al, 2020), used for productivity analysis purposes.…”

Section: Construction Productivitymentioning

confidence: 99%

Integration of deep learning and extended reality technologies in construction engineering and management: a mixed review method

Lee

Rashidi

Talei

et al. 2022

View full text Add to dashboard Cite

Purpose In recent years, deep learning and extended reality (XR) technologies have gained popularity in the built environment, especially in construction engineering and management. A significant amount of research efforts has been thus dedicated to the automation of construction-related activities and visualization of the construction process. The purpose of this study is to investigate potential research opportunities in the integration of deep learning and XR technologies in construction engineering and management. Design/methodology/approach This study presents a literature review of 164 research articles published in Scopus from 2006 to 2021, based on strict data acquisition criteria. A mixed review method, consisting of a scientometric analysis and systematic review, is conducted in this study to identify research gaps and propose future research directions. Findings The proposed research directions can be categorized into four areas, including realism of training simulations; integration of visual and audio-based classification; automated hazard detection in head-mounted displays (HMDs); and context awareness in HMDs. Originality/value This study contributes to the body of knowledge by identifying the necessity of integrating deep learning and XR technologies in facilitating the construction engineering and management process.

show abstract

Section: Construction Productivitymentioning

confidence: 99%

Integration of deep learning and extended reality technologies in construction engineering and management: a mixed review method

Lee

Rashidi

Talei

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Some exceptions are the work by Jung et al [4], using movies of construction equipment downloaded from YouTube and two papers [5,6] which used unmanned aerial vehicles (UAV) to collect vision-based data without the constraint of fixed camera positions. Additionally, an early work by Vachkov et al [3] and a recent work by Saari and Odelius [7] applied unsupervised learning techniques for MAR; whilst some research works started to use simulated data [8,9] and augment the data with known invariances [10]. However, apart from these research works, the MAR work has remained fully within the supervised learning paradigm, without using unlabeled data or synthetic data for the learning process.…”

Section: Literature Review 21 Machine Activity Recognitionmentioning

confidence: 99%

Semi-Supervised Learning for Forklift Activity Recognition from Controller Area Network (CAN) Signals

Chen

Rögnvaldsson

Nowaczyk

et al. 2022

Sensors

View full text Add to dashboard Cite

Machine Activity Recognition (MAR) can be used to monitor manufacturing processes and find bottlenecks and potential for improvement in production. Several interesting results on MAR techniques have been produced in the last decade, but mostly on construction equipment. Forklift trucks, which are ubiquitous and highly important industrial machines, have been missing from the MAR research. This paper presents a data-driven method for forklift activity recognition that uses Controller Area Network (CAN) signals and semi-supervised learning (SSL). The SSL enables the utilization of large quantities of unlabeled operation data to build better classifiers; after a two-step post-processing, the recognition results achieve balanced accuracy of 88% for driving activities and 95% for load-handling activities on a hold-out data set. In terms of the Matthews correlation coefficient for five activity classes, the final score is 0.82, which is equal to the recognition results of two non-domain experts who use videos of the activities. A particular success is that context can be used to capture the transport of small weight loads that are not detected by the forklift’s built-in weight sensor.

show abstract

Action Recognition of Excavator with Data Augmentation of Simulator-Generated Training Data

Kasahara¹,

Sim²,

KOMATSU³

et al. 2022

Journal of the Japan Society for Precision Engineering

View full text Add to dashboard Cite

Action Recognition of Construction Machinery from Simulated Training Data Using Video Filters

Cited by 3 publications

References 0 publications

Integration of deep learning and extended reality technologies in construction engineering and management: a mixed review method

Integration of deep learning and extended reality technologies in construction engineering and management: a mixed review method

Semi-Supervised Learning for Forklift Activity Recognition from Controller Area Network (CAN) Signals

Action Recognition of Excavator with Data Augmentation of Simulator-Generated Training Data

Contact Info

Product

Resources

About