A modified-YOLO based vehicle-pedestrian detection model for improved urban mobility.

Congestion on traffic lanes is a key issue impeding the growth of a metropolitan metropolis. The reason for this is the growing number of cars on the road, which causes significant time delays at traffic lights. Several strategies and procedures have been developed over the years to address this issue and make traffic control systems more dynamic. The static traffic control systems operated on predetermined timings that were assigned to each traffic lane and could not be changed. There was also no mechanism for counting and detecting pedestrians at zebra crossings, nor for detecting emergency vehicles in traffic. Here in this research, we will investigate multiple machine learning and deep learning models for car and pedestrian identification in traffic-prone lanes. Furthermore, we will propose a system that will contain a deep learning model to recognize various types of cars, pedestrians, and overlapping and intersecting vehicles on traffic lanes. Finally, a choice will be taken to free up the crowded lanes depending on the traffic count on each lane. [ABSTRACT FROM AUTHOR]