Enhanced Object Detection using YOLOv8: Identifying Vehicles and Pedestrians in Urban Environments

With the vast usage of cameras and sensors for vehicles and crowd management in smart cities, their detection and tracking are paramount. Traditional cars and pedestrian detection methods have limitations in the motion of objects, accuracy, detection in real-time, etc. This paper uses a Vehicles and Pedestrian detection approach based on the YOLOv8 algorithm, which leverages the advantages of deep networks for real-time object detection and mitigation of false alarms, occlusion handling, and motion blur. The targeting classes such as cars, pedestrians, cyclists, trucks, vans, trams, and people sitting were used for experimentation and validation of the approach. YOLOv8's architecture was deliberately chosen for its optimal balance of speed and accuracy. The paper outlines the dataset preparation, training, and evaluation processes, emphasizing adaptations for diverse conditions. YOLOv8's architecture was deliberately chosen for its optimal balance of speed and accuracy.