Low-Cost IoT and Machine Learning System for Secure Parking Access

Ramirez M.; Landeo F.; Wong L. · International Conference on Electrical, Computer, and Energy Technologies, ICECET 2026 (2026) · DOI: 10.1109/icecet65726.2026.11632951

Resumen

Security at access to vehicle parking is a recurring problem in Peru due to the lack of automated systems that allow the identity of drivers and vehicles to be reliably validated. This situation generates vulnerabilities and limits operational control of parking lots. The objective of this work is to build a low-cost technological solution that improves vehicle entry security through the use of Machine Learning and IoT. The methodology uses a five-layer architecture ('Capture', 'Communication', 'Business Logic', 'Storage', 'Presentation'), which will integrate ESP32-CAM cameras, facial recognition, vehicle license plate detection using the YoloV5 Algorithm, EasyOCR and communication via MQTT. The tests were carried out in a controlled environment, where a driver and his vehicle participated, and the authentication of face and license plate was evident, which allowed safer access, demonstrating the viability of the proposed system.

Instituciones: Universidad Peruana de Ciencias Aplicadas