2026, Vol. 13, No. 2. - go to content...

Permanent address of this page - https://t-s.today/en/12sats226.html

Метаданные этой статьи так же доступны на русском языке

DOI: 10.15862/12SATS226 (https://doi.org/10.15862/12SATS226)

Full article in PDF format (file size: 958.5 KB)

For citation:

Reshetnikova A.V., Razov I.O. Control of construction and installation works through computer vision technology. Russian Journal of Transport Engineering. 2026; 13(2). Available at: https://t-s.today/PDF/12SATS226.pdf (in Russian). DOI: 10.15862/12SATS226

Control of construction and installation works through computer vision technology

Reshetnikova Aleksandra Vladimirovna
Industrial University of Tyumen, Tyumen, Russia
Construction Institute
E-mail: Reshetnikova389@yandex.ru
ORCID: https://orcid.org/0009-0004-1578-0326
RSCI: https://elibrary.ru/author_profile.asp?id=1336923

Razov Igor Olegovich
Industrial University of Tyumen, Tyumen, Russia
Construction Institute
E-mail: razovio@mail.ru
ORCID: https://orcid.org/0000-0002-4829-7080
SCOPUS: https://www.scopus.com/authid/detail.url?authorId=57190863789

Abstract. The article examines the problem of maintaining a high level of occupational injuries in the construction industry of the Russian Federation, where the death rate is 2.58 times higher than the national average. The relevance of the study is due to the limitations of traditional visual surveillance methods and the lack of scientifically sound approaches to the multitasking application of computer vision on construction sites. The aim of the work was to theoretically substantiate the possibility of integrated use of computer vision systems for simultaneous monitoring of the quality of construction and installation work, compliance with labor protection requirements and logistical optimization within the framework of a single information environment of the facility. The authors conducted an analysis of existing publications on the topic, which revealed the fragmentary nature of research: most of the work is aimed at solving isolated tasks — either detecting structural defects, monitoring the use of personal protective equipment, or tracking the movement of equipment. The article consistently examines the technical stages of visual data processing, including noise pre-filtering, neural network image segmentation, and justifies the use of the infrared spectrum to overcome interference caused by dust and precipitation. The prospects of integrating computer vision with information modeling of buildings and the Internet of Things to create a «smart» construction site are shown. The functional capabilities of quality control are systematized: identification and classification of defects, automated accounting of structures and tools by unique identifiers. The potential of round-the-clock monitoring of compliance with safety regulations and optimization of logistics, including with the use of unmanned aerial vehicles, has been revealed. The requirements for system performance have been defined, including an acceptable delay in transmitting a violation signal of up to 50 milliseconds. The advantages of an integrated approach are systematized and the limitations associated with the sensitivity of algorithms to meteorological factors are outlined.

Keywords: computer vision; construction industry; occupational injuries; quality control; safety technology; building information modeling; Internet of Things