Computer Vision Based Enhanced Traffic Management System Using YOLO

Traffic congestion is deteriorating day by day. The traditional traditional traffic signals cannot deal with the ever- increasing number of vehicles appropriately. We have combined computer vision and machine learning to replicate the complicated inflow and outflow of traffic in signalized crossings. You Only Look Once (YOLO) is a state- of- the- art object identification system based on deep convolutional neural network, which we have used in this context. This research proposes the need to implement this method properly to maximize the number of cars that may safely cross while minimizing waiting time. In this method, the traffic signal phase is mostly determined by the collected data which includes waiting time per vehicle and congestion density. YOLO can be implemented using transfer learning algorithms in the embedded controllers. Vehicle counting is the estimation of traffic density for estimating traffic conditions. With their extensive use in urban traffic, video surveillance today forms an integral part of the Central data source. Recently, real- time traffic management systems have been developed using handheld and mobile cameras and big data processing. Here we propose a vehicle counting approach based on video frames captured by a handheld traffic camera. The video processing is divided into three steps: object recognition using YOLO, tracking using a correlation filter, and finally counting. Regarding the number of persons, the findings from the correlation and YOLO filters are extremely accurate. For example, correlation filters are utilized to attain improved accuracy and comparable tracking speed, even if the YOLO model performs better in detection than the others. Therefore, a multi- object tracker based on correlation filters is constructed using the YOLO model's result bounding boxes. Empirical research using real- world video recordings demonstrates that the suggested approach is capable of detecting, tracking, and measuring moving vehicles.