A Research on Pedestrian-Vehicle Detection Algorithm based on YOLO

Abstract

With the continuous advancement of urbanization, the importance of pedestrian and vehicle detection is becoming more and more prominent in many fields such as intelligent transportation, autonomous driving, and public safety. Traditional methods of pedestrian-vehicle detection often suffer from slow speed and inaccuracy. In this paper, we systematically review and deeply analyze target detection algorithms, with a focus on the progression from YOLOv1 to YOLOv10 in pedestrian-vehicle detection, and summarize the various network enhancement strategies reported in the literature. Despite progress in feature extraction, network structure, algorithm optimization, lightweight design, and multimodal exploration, pedestrian-vehicle detection algorithms still confront significant challenges, particularly in detecting extremely small targets and adapting to various environments. This paper offers a thorough perspective and serves as a valuable reference for research in pedestrian-vehicle detection, significantly advancing the development of more precise and resilient detection technology.