Desarrollo de un sistema de identificación de personas con un drone mediante el uso de redes neuronales artificiales y visión artificial

Abstract

The use of facial recognition systems represents an important advance in the area of artificial vision, this technological field evolves day by day, which is why different applications are frequently exhibited based on this technology that can be applied in various sectors to provide security and control. The present work develops a system of identification of people with a drone through the use of artificial networks and artificial vision which uses the Python programming language, based on TensorFlow and Keras for the creation of the artificial neural network and HaarCascade as a method. face detection whose code is stored in a Lenovo Laptop that will be connected to the UAV. What will allow the acquisition, the pre-processing of images, the comparison and digital processing of the samples for the learning of the neural network and in this way the identification of the wanted persons. This type of system has an educational and demonstrative approach for new students who study the subjects of intelligent systems and artificial intelligence at the Salesian Polytechnic University, Quito-South Campus. The purpose of this work focuses on the field of security and control of people within the university area. Through the implementation of the system, both the neural network and the artificial vision, its effectiveness was verified in a controlled environment, resulting in an effectiveness of 80% in the detection of the face of the person to be identified.