Helmet detection is a computer vision task with important application value, involving safety management in many fields such as construction sites, mines, and electric power. However, helmet detection also faces many challenges, such as large changes in target size and aspect ratio, rapid changes in target velocity, target occlusion, and background interference. In order to solve these problems, this paper proposes a safety helmet detection method based on YOLOv8, which uses the characteristics of high speed and high precision of YOLOv8 combined with the characteristics of safety helmets to achieve effective detection and identification of safety helmets.