Vehicle Security Systems using Face Recognition based on Internet of Things

摘要

Nowadays, the automobile sector is one of the hottest applications, where vehicles can be intelligent by using IoT technology. But unfortunately, these vehicles suffer from many crimes. Hence it has become a big challenge for the IoT to avoid such these crimes from professional thieves. This paper presents a proposal for the development of a vehicle guard and alarm system using biometric authentication based on IoT technology. Whereas, for vehicle security issues; the proposed system VSS ? IoT gives only full access for authorized vehicle’s driver based on the interface of a Raspberry Pi 3 Model B + development board, Pi camera, PIR sensor, and smart-phone. Therefore, if the proposed system detects an unauthorized person inside the vehicle, then the system will notify and send his image to vehicle’s owner and/or to a police workstation through the Internet, as well as, its location in case the vehicle is stolen or damaged. The proposed system is tested on two datasets that are ORL dataset and our dataset. The experimental results of the VSS ? IoT showed that the accuracy is 98.2% on ORL dataset, whereas 99.6% when applied on our dataset. Besides, the V SS ? IoT enhances the sensitivity to 97.7% which is important for real-time. As well as the result demonstrated that the proposed system took shorter time 0.152 sec under different illumination conditions, when the value of the threshold is 3 * 10~(3)and 3.50 * 10~(3). Therefore, the VSS ? IoT is very robust and reliable for face recognition when deployed on the low-power processor.