Computer Vision for Industrial Safety and Productivity

Abstract

The growing deployment of computer vision in industrial processes significantly contributes to strengthening the manufacturing sector in terms of productivity and safety of the workers. Manufacturing workers are often working in hazardous environments handling different dangerous equipment putting their life on the line every day. Work accidents are reminders for which companies must make efforts to reduce its occurrence and their adverse impact on the lives of workers. In case of an active accident, the computer vision system can send an alert to managers and staff about location and the intensity of the accident so the production process can be halted in that specific area and proactively ensure the safety of employees. The deployment of computer vision-powered systems operating 24/7 accelerates manufacturing cycles increasing productivity. Computer vision applications have a major role in product and component assembly in the manufacturing space. They also aid in defect detection with increased accuracy and precision. Manufacturers conduct constant monitoring of equipment used for production manually. To improve the safety and working conditions for the workers and increase productivity in the manufacturing sector, this project aims to implement computer vision as a monitoring method to assure the security measures are followed and analyze the productivity in the organization. The object recognition algorithm, YOLOv3, is trained and tested using data that is gathered from industrial facilities in the form of images.