It is very critical within the scope of occupational health and safety that employees use the necessary personal protective equipment (PPE) in workplaces such as construction, mines, factories, warehouses and shipyards. Not using PPE puts employees at great risk. In this study, various object detection models are trained in order to continuously monitor employee safety through IP cameras in the workplace for various industries. An advanced AI-powered image labeling tool has been developed and used, which provides data security standards, has various data augmentation options and has auxiliary model support in the labeling process. With the developed tool, a data set was obtained by marking the images containing various personal protective equipments and people that are obtained from the internet in the context of occupational health and safety, and this data set was expanded with data augmentation. By trying different methods, up to 95.3% success was achieved.