Smart Traffic Management: A Deep Learning Approach for Congestion Reduction using YOLOV8 & RandomForest

Traffic management has become a tedious task for policemen to carry out manually. According to the publication produced by Bureau of Police Research & Department in 2017, India has just about 72,000 traffic cops to manage almost $20,000,000$ vehicles. As the ratio is insanely high and unmanageable, traffic congestion has increased ever since. Hence, we proposed a smart traffic management system which will particularly tackle this problem using cutting edgetechnology. By analyzing the oncoming traffic, our model predicts the required time for particular vehicles to pass and adjusts the signal timings accordingly. By combining models from deep learning and machine learning we, propose a cheap, efficient system that addresses these problems by a combination of YOLOv8 & Random Forest Regression to generate results. These results consist of annotated images and timings for signals which are calculated from the vehicle count.