Implementación de un sistema de seguridad automotriz basado en visión artificial para evitar robos del computador.

Abstract

The objective of this graduate research was to implement an automotive security system based on an artificial vision to prevent theft of the computer automotive. Composed of a set of sensors that were located in the four doors of the vehicle and the computer automotive, in addition, a camera capable of obtaining a digital collection of images was used in real-time that was located in the internal part of the vehicle according to the requirements of the implementation, to guarantee the safety of the auto part and the owner peace of mind. The Arduino board was in charge of processing the information collected by the sensors. In turn, communication was established with the Raspberry Pi for image recognition captured in real- time, compared with a record of inscription images identifying the individual's identity for later storage. The Global Positioning System (GPS) obtains the vehicle's location, and the safety alerts are sent through the Global System for Mobile communications (GSM) by messages text to the owner's mobile phone. The algorithm was made in Python for its diversity and compatibility with libraries dedicated to face recognition. also, it was used OpenCV to have a camera control. As a result of the tests, it was determined that the adequate distance for a correct detection must be greater than 30 cm from the steering wheel to the right up to 100 cm as the maximum distance and the speed that the system took to recognize the user is 17.75s. It is recommended to extend the prototype testing time in order to identify improvements in the design.