Automatic traffic surveillance system Utilizing object detection and image processing

摘要

In our day and age where the numbers of cars on the road are rapidly increasing, thereby causing traffic. Drivers are becoming more reckless and carefree as the burden on the current human and automated system grows. Drivers and bikers who may wish to save a few minutes may break red lights and avoid wearing helmets but these small actions can have a significant impact and can result in the loss of lives. We propose a system that will intelligently use deep learning-based object detection to identify traffic offenders and provide methods to penalize them by recognizing their number plate. Our system will be able to detect traffic light violators and bikers without helmets. It has been designed in such a way that it is robust enough to work in drastic conditions and intelligent enough to reduce human dependence.