Implementación de un sistema de seguridad para supervisión de niños entre 2 a 4 años usando Visión Artificial.

Abstract

Due to the large number of accidents that have been registered in homes, a security system was implemented for children from 2 to 4 years using artificial vision, using elements such as Raspberry Pi 3. Arduino UNO. SIM module 900. DF Player. relay of a channel, screen, beacon, speakers among others for the construction of the prototype. The development of the detection system was based on the algorithm called "Haar Cascade", the creation of the interface was made using the graphical user interface (GUI) Cascade Trainer. Based on this, the system was developed to detect dangerous objects for children. When the system detects a dangerous object such as scissors, knives, etc. activates an emergency alarm that immediately issues a call to the number registered in the database. The Cascade algorithm allowed to create a database of positive and negative images in this case the positive images will be the objects to be detected. The graphical interface is visualized by the user allowing to manipulate six buttons, each one fulfils specific functions that are: start camera, pause, detection, activate alarm, reboot and test GSM. According to the tests carried out, it is established that the system works optimally in controlled spaces, while in uncontrolled spaces the parameters vary by 0.38% the response time of the system, in 2% the number of objects recognized, in 1% the distance of the objects to be detected and 2% the call response time, that is, the system works optimally, but it takes time to send the answers.