Deep Learning & Computer Vision for IoT based Intelligent Driver Assistant System

With the exponential increment of vehicles, roadside accidents also have been increasing rapidly where approximately 80% of these accidents were caused by human error. Therefore, the automobile industry and the government authorities are more focused on accident prevention by introducing improved road safety systems for the general public. Driver assistance system (DAS) is an intelligent road safety development where it senses the surrounding of a moving vehicle which will assist the driver to avoid the dangers and also warn the drivers of immediate dangers. With the current technological advancements, the automobile industry is equipped with the internet of things (IoT) based data transfer mechanisms, wherewith the concept of ‘connected car’ the passengers and the other interconnected vehicles with the internet can share data with back end applications. The data is consisting of the current location, the distance travelled by the vehicle, whether the vehicle requires urgent service. This study is mainly focused on the development of an intelligent driver assistance system based on computer vision and deep learning, which can prevent accidents with early detection of drowsiness, harmful objects and also by alerting the driver with the signboards and the road lines. The system is capable of passing the emergency messages to the driver as well as the other interconnected vehicles through the website by communicating the real-time road map generated within the system. The proposed system has been implemented and tested with multiple detection scenarios where machine learning has been employed to improve the accuracy of the results.