Implementación de un Control de Acceso con Reconocimiento Facial Mediante Opencv para la Sala de Profesores de la FIE.

Abstract

The access control system using facial recognition for staff roomof the faculty of Computer Science and Electronics of the Polytechnic School of Chimborazo. It is designed so that we can apply new technologies to improve income security staff. Applying the deductive method of facial recognition system based on a statistical method that calculates the values and vectors of the covariance matrix of color intensity in the gray space. The system grants Access to the person, if this is within the set time by the administrator also provides a personal password in case of failure in the face detection, if the user is in the established schedule is driven by the output relay system. The system consists of a camera (raspicam) which captures the image of each stage of the system, an alphanumeric keypad is displayed to enter the different data for each teacher, the entire system is controlled by a computer reduced plate (raspberry) which processes the images and data of each person. Through three experimental tests it was posible to determine an efficiency rate of 78%, which is acceptable in systems existing face recognition. It was concluded that the library Eigenfaces machine visión library (OpenCV) proved to be a suitable tool, allowing to have a quick confrontation and a high percentage of reliability in recognizing the image of each teacher. It is recommended to incorpórate a light source (60W), because without proper lighting the image monocromiza and fails in the face recognition system.