The surveillance system detects in real-time whether a worker is wearing a safety helmet and the medical mask required to prevent the spread of COVID-19.
About 30 cameras are deployed at production sites and where employees gather, with the system also capable of detecting improper use of safety clothing and violations of the use of PPE and workers involved in operations in vehicles.
"A pilot test of the system launched in March 2020 reduced the number of violations by almost three times," said Ilya Sterlikov, director of labour protection and industrial safety. "This system is an important step towards reducing disease and improving overall health and safety at work."
The use of neural networks in the video surveillance system has led to an 84% accuracy rate, said the company.
Automated reports of violations are sent to the labour protection and industrial safety department for human checks, with operators then providing feedback to heads of departments.