AUTOMOTIVE THEFT DETECTION USING FACE RECOGNITION

Shejina V, Asil A

Abstract: Nowadays vehicles are becoming an important part of human life. Vehicles have a tremendous role on lightening the life styles of people. At the same time the rate of vehicle theft is increasing day by day. Here comes the importance of detection of vehicle theft. The existing security system is based on anti theft alarm. Whenever a person touch the body of the vehicle , the alarm is automatically turned on and sounds. The limitation of the existing system is that an intruder can break the security alarm when he is able to access the power source. The aim is to develop a security system which can be incorporated inside a vehicle for the purpose of detecting theft. The proposed system is a face recognition based security system with face recognition module loaded in Raspberry Pi . When the driver sits in the driver seat and inserts key in to the key hole, a sensor is activated and a hidden camera captures the image of the driver which is then sent to the Raspberry Pi through an RS-232 cable. The captured image is pre-processed using histogram equalization .The pre-processed image is given for face detection. Face detection is performed using Viola-Jones algorithm. The extracted face is recognized using the Local Binary Patterns Histograms algorithm. If the vehicle is stolen, the software sends an mms to the owner comprising the image of the driver and the current location of the vehicle. The system is well implemented in OpenCV- Python and trained to identify two possible drivers. The software can distinguish between authorized and unauthorized drivers and send email to the owner if the vehicle is driven by an unauthorized person.

Keywords: Raspberry Pi B+, Histogram Equalization, Viola-Jones algorithm, Local Binary Patterns Histogram

DOI: https://doi.org/10.15623/ijret.2016.0531009