Sistema de detección de situaciones delictivas en establecimientos comerciales usando inteligencia artificial

Abstract

The constant increase in insecurity in Ecuador is a phenomenon that manifests annually through a variety of crimes, with armed robberies in commercial establishments being the main focus of the present research. In this context, a study was carried out with the purpose of developing an artificial intelligence-based system for the detection of criminal situations in commercial premises. The primary objective of this system is to generate alerts in the face of potential assaults, with the aim of expediting the response from the relevant authorities. This system was built in three stages: data acquisition, processing, and visualization. In the acquisition phase, a DS-2CD2147G2 IP camera with 5 megapixels is used to capture real-time images. The captured data is transmitted to the NVIDIA JETSON NANO microcomputer for processing. The system was developed in Python, and Yolov8 was chosen as the artificial vision algorithm, responsible for processing the data and recognizing indicators associated with armed robbery. The visualization of the results and generated alerts is sent through Telegram, providing images of the crime, a detailed description of the situation, and the location. It is relevant to highlight that the information is processed and transmitted in real-time, ensuring an efficient response to potential criminal incidents. The system achieved an effectiveness of 84% after conducting the corresponding tests.