Tesis Telecomunicaciones

Permanent URI for this collectionhttp://repositorio.uta.edu.ec/handle/123456789/34848

Browse

Filter results by typing the first few letters
Now showing 1 - 1 of 1
  • No Thumbnail Available
    Item
    Sistema CCTV con alerta temprana y reconocimiento de armas de fuego utilizando machine learning
    (Universidad Técnica de Ambato. Facultad de Ingeniería en Sistemas, Electrónica e Industrial. Carrera de Telecomunicaciones, 2025-02) Unapucha Unapucha Jorge Miguel; Ayala Baño Elizabeth Paulina
    The continuous increase in insecurity in Ecuador is reflected annually through various criminal activities, with armed assaults and kidnappings in homes being the main focus of this analysis. In this context, a study was conducted aimed at designing a system based on machine learning techniques for the identification of firearms used in home assaults. The primary goal of the system is to generate early alerts to expedite the intervention of the relevant authorities. The system comprises three stages: acquisition, processing, and visualization. In the data acquisition stage, a Closed-Circuit Televisión consisting of two 2MP cameras and a Hilook brand Digital Video Recorder is used to capture frames in real time. The captured data is sent to the NVIDIA JETSON NANO microcomputer for further processing. The system was developed in Python, and Yolov5 was selected as the artificial vision model responsible for processing the acquired frames, resulting in the identification of firearms in home assaults or kidnappings. The visualization of the identified elements and the generated alerts is carried out through phone calls and spam messages, providing a detailed description of the situation and the identified object, along with an integrated audible alert. It's important to highlight that the data is analyzed and transmitted immediately, ensuring an fast response to potential criminal acts. As a result of the tests conducted, the system achieved an effectiveness of 85%.