Proposing a route recommendation algorithm for vehicles based on receiving video

Abstrak

In this paper, we propose a method to classify traffic status for the route recommendation system based on received videos. The system will determine the number of vehicles in the region of interest (ROI) to determine and calculate the coefficient of variation (CV) based on the videos extracted from cameras at intersections. It then predicts the congested traffic junctions in the city. The data then goes through the routing module and is transmitted to the website to find the best path between the source and destination requested by users. In this system, we use you only look once (YOLOv5) for vehicle detection and the A* algorithm for routing. The results show that the proposed system achieves 91.67% accuracy in detecting traffic status comparing with YOLOv1, deep convolutional neural network (DCNN), convolutional neural network (CNN), and support vector machine (SVM) models as 91.2%, 90.2%, 89.5%, and 85.0%, respectively.