Facial recognition system for employee identification, action logging and access control.
The customer is a company from Russia, which is engaged in processing and production of food products.
It was necessary to develop a system of identification of employees of the enterprise by means of CCTV, to introduce personal logging of employees (actions, movements around the plant, etc.) and to integrate the facial recognition system into the existing system of access control to certain production areas.
The application is implemented as a desktop application in C+ and Python, QT was chosen as the main framework for UI implementation. OpenCV and YOLO were taken as the basis for identification of faces and actions of employees. To optimize the system performance, several face recognition algorithms were used: a simpler model (83% recognition accuracy) not requiring large computing power was used to identify employee movements around the enterprise in real time, more accurate algorithms with recognition accuracy up to 98% were used to control access to production sectors. Personalized logging of employee actions using YOLO with linking to timestamp video was implemented.
Scrum was used to manage the development of the project using Agile methodology, which allowed to get a working prototype in the shortest possible time and gradually increase the functionality to the required for the customer.
The developed system made it possible to monitor the actions of employees in order to exclude industrial injuries and cases of negligence in production, as well as for prompt internal investigation and elimination of violations of production techniques, which in turn should improve the quality of manufactured products.
