ACCIDENTAL DOORING WARNING AND ASSIST SYSTEM

The term dooring refers to the accidents where a vehicle door is suddenly opened into the path of an oncoming cyclist or motorcyclist. These issues are serious road safety hazards, especially in congested urban areas where people may be careless. This project proposes an intelligent, computer vision based warning system designed to detect the approaching vehicles and alert the occupants in the vehicle before opening the door. The system uses the rear parking camera in vehicles with rear parking system or a USB webcam in vehicles without rear parking system to capture live video feed, which is processed by a Raspberry Pi 4 microprocessor. The system will run a lightweight deep learning model, MobileNetSSD, through TensorFlow Lite. Detected objects are classified in real time, and the relative distance of other vehicles is estimated using bounding box dimensions. If a vehicle is identified within a preset safety distance threshold, a clear visual warning is displayed to prevent door opening in unsafe conditions. The system was successfully implemented and tested on both a laptop and Raspberry Pi, offering reliable detection and timely warnings with minimal hardware requirements. This solution offers a cost-effective and scalable approach to vehicle door safety, with potential future enhancements including automatic door locking, audio alerts, and integration into commercial driver-assistance systems (ADAS).