Car Theft Detection Using Face Recognition

Abstract

Vehicle theft is a growing concern due to the increasing number of unauthorized vehicle access incidents. Conventional vehicle security systems such as alarms and key-based locks are often insufficient against modern theft techniques. This paper presents a smart vehicle security system using ESP32-CAM and Telegram for real-time monitoring and owner-based authorization. When an attempt is made to start the vehicle, the system captures an image using an onboard camera and sends it to the vehicle owner through a Telegram bot along with the vehicle’s live GPS location. The vehicle ignition is enabled only if the owner approves the request. If the request is rejected or suspicious activity is detected, the system activates an alarm and sends theft alerts to the owner. The proposed system is lightweight, event-driven, and does not rely on complex machine learning models, making it suitable for real-time embedded applications. Experimental implementation shows that the system effectively enhances vehicle security with fast response time and minimal hardware requirements.