IoT-based Automated Theft Detection in Vehicles: A Combined Approach of YOLO Detection and MediaPipe Pose Analysis

Vehicle theft has emerged as a notable societal concern within the United States. Despite advancements in detection technology, rates of vehicle theft persist in an upward trend. In response to this exigent issue, an Internet of Things (IoT) system has been devised to confront the challenges encountered by car owners, particularly those whose vehicles lack inherent theft prevention mechanisms. The objective of this research is to introduce a comprehensive system for preventing vehicle theft. The proposed system integrates two fundamental technologies: computer vision for identifying suspicious activity, and vibration and pressure sensors to detect potential physical breaches into the vehicle. Upon detection of suspicious activity, the proposed system quickly dispatches a notification to the user along with a 20-second video capturing the behavior. This IoT system surmounts the constraints of conventional anti-theft systems, thereby enhancing the security of vehicle owners in the United States and curbing the proliferation of vehicle theft.