Human Detection Based on YOLOv5 Network

Abstract

With the vigorous development of the economy, the improvement of people's quality of life, the continuous increase in the ownership rate of family cars, and the increasing pressure on traffic management, people urgently hope to use new technologies to improve and optimize the past system. Therefore, the research on the modernization and intelligence of traffic control has become a research hotspot for scholars. When the crowd density is too high, all kinds of accidents are easy to happen, and it will bring difficulties to the management of order managers. Timely detection of crowd density, and when the crowd density is too high, warning information will be issued, which is conducive to the managers to evacuate the crowd in time. So, the final project is about how to recognize humans in the image, and with men and women. Today's human detection can be broadly divided into traditional image processing methods and machine learning-based methods. The traditional image processing method detects the human by preprocessing the image and then detecting the human contour. Frequent stampede events in large events at home and abroad have caused a lot of casualties. Therefore, the research on crowd counting has attracted more and more attention. If the crowd density of the current scene can be accurately estimated and the corresponding security measures can be arranged, the occurrence of such incidents can be effectively reduced or avoided. Crowd counting is widely used in video surveillance, traffic monitoring, public safety, urban planning, and the construction of intelligent shopping malls.