An issue of non-compliance with occupational safety requirements for using safety helmets in production is discussed, considering that the traditional control method used by managers is inconsistent and inefficient. An automated control system based on computer vision algorithms is proposed as a solution.
The goal of this study is to develop a computer vision system and evaluate its efficiency in real production conditions using key metrics, i.e., the accuracy of identification, the number of violations, and its impact on industrial discipline. The study's methodology includes data collection and comparative analysis before and after the system’s implementation, utilizing video recording tools and subsequent verification of results. The test results have shown that the accuracy and completeness of recognizing the absence of a helmet under variable lighting conditions and various spatial positions of an employee were equal to 98,9 % and 97,8 %, respectively. The following pattern has been identified during the violation dynamics analysis: in the first week after the implementation, a significant number of violations associated with the non-use of personal protective equipment by employees were registered. Following the imposition of disciplinary sanctions for violators, a significant drop in such violations was observed. The data obtained indicates a stable behavior pattern as a result of the realization of constant control by the employees. This leads to the conclusion that digitalization establishes a feedback cycle, where the automated recording of violations and the implementation of administrative measures contribute to the formation of a culture of compliance with occupational safety requirements.
Karchina E.I., Ivanova M.V., Glebova E.V. Risk-Oriented Approach to Providing Employees with Personal Protective Equipment. Bezopasnost truda v promyshlennosti = Occupational Safety in Industry. 2022. № 9. pp. 84–90. (In Russ.). DOI: 10.24000/0409-2961-2022-9-84-91
2. Jocher G., Chaurasia A., Qiu J. Ultralytics YOLO.. 2023. Available at: https://github.com/ultralytics/ultralytics (accessed: December 12, 2025).
3. Причины производственного травматизма в России за 2024 год. URL: https://ot-media.ru/hse/analytics/prichiny-proizvodstvennogo-travmatizma-v-rossii (дата обращения: 12.12.2025).
Causes of industrial injuries in Russia for 2024. Available at: https://ot-media.ru/hse/analytics/prichiny-proizvodstvennogo-travmatizma-v-rossii (accessed: December 12, 2025). (In Russ.).
4. Чумаков К.В. Тенденции в современных системах мониторинга промышленной безопасности и охраны труда // Известия Томского политехнического университета. Промышленная кибернетика. 2024. Т. 2. № 4. С. 34–40.
Chumakov K.V. Application of video analytics to monitor and improve the safety of technological processes. Izvestiya Tomskogo politekhnicheskogo universiteta. Promyshlennaya kibernetika = Bulletin of The Tomsk Polytechnic University. Industrial cybernetics. 2024. Vol. 2. № 4. pp. 34–40. (In Russ.).
5. Шарманов В.В., Симанкина Т.Л., Горбачев И.А. Идентификация местоположения работника на объекте строительства с помощью цифрового двойника // Вестник научного центра по безопасности работ в угольной промышленности. 2022. № 3. С. 70–80.
Sharmanov V.V., Simankina T.L., Gorbachev I.A. Identification of a Workers Location at a Construction Site Using a Digital Twin. Vestnik nauchnogo centra po bezopasnosti rabot v ugolnoj promyshlennosti = Bulletin of the scientific center for safe operation in the coal industry. 2022. № 3. pp. 70–80. (In Russ.).
6. Попова Ю.Б. От LMS к адаптивным обучающим системам // Системный анализ и прикладная информатика. 2019. № 2. С. 58–64. DOI: 10.21122/2309-4923-2019-2-58-64
Popova Yu.B. From LMS to adaptive training systems. Sistemnyj analiz i prikladnaya informatika = System analysis and applied information science. 2019. № 2. pp. 58–64. (In Russ.). DOI: 10.21122/2309-4923-2019-2-58-64
7. Безопасность RFID-систем / В.С. Бельский, Е.С. Грибоедова, К.Д. Царегородцев, А.А. Чичаева // International Journal of Open Information Technologies. 2021. T. 9. № 9. C. 1–20.
Belskij V.S., Griboedova E.S., Caregorodcev K.D., Chichaeva A.A. Security of RFID systems. International Journal of Open Information Technologies. 2021. Vol. 9. № 9. pp. 1–20. (In Russ.).
8. Пильцов М.В. Разработка систем сбора данных на основе платформы ARDUINO и языка программирования PYTHON // Современные технологии и научно-технический прогресс. 2018. № 1. С. 84–85.
Pilcov M.V. Developing data collection systems based on the ARDUINO platform and the PYTHON programming language. Sovremennye tekhnologii i nauchno-tekhnicheskij progress = Modern technologies and scientific and technological progress. 2018. № 1. pp. 84–85. (In Russ.).
9. Викторов М.Ю. Цифровые двойники. Основные аспекты внедрения и применения технологий информационного моделирования (ТИМ) в строительстве // Журнал «Русский инженер». 2021. С. 15–16.
Viktorov M.Yu. Digital twins. Basic aspects of the implementation and use of information modeling technologies in construction. Zhurnal «Russkij inzhener» = “Russian engineer” journal. 2021. pp. 15–16. (In Russ.).
10. Mühlbauer M., Kutzner K., Sommer A., Würschinger H., Hanenkamp N. An Approach to Progress Monitoring of Industrial Manual Processes Based on Camera Recordings and Object Interactions. Procedia CIRP. 2022. Vol. 107. pp. 582–587. DOI: 10.1016/j.procir.2022.05.029