Intelligent System Detecting Safety Requirements Violation in Railway Infrastructure Maintenance

Safety improvement issues concerning workers engaged in the repair and maintenance of railway infrastructure are considered in this paper. The main purpose of the work is to improve labour protection requirements compliance through the development and implementation of an intelligent system for detecting violations of safety requirements during work on railway infrastructure. The object of the research is to develop detection violation algorithms for labor protection of railway workers. As part of the research presented, the data necessary for the system learning models was collected and marked up. A generalized algorithm for detecting violations in using personal protective equipment has been developed based on the work team detection model and the violation classification of using personal protective equipment. Functioning of the detection violation model in terms of labour protection requirements under different methods of image interpolation has been researched. The research showed that the number of training epochs and the type of interpolation affect the quality of the model tested. The research presented logically integrates into the semantic network of predictive analytics models in railway transport and confirms the practical value of the previously obtained theories on the construction of such model hierarchy.