Recognition of children’s movement patterns during recess for characterizing particle exposure doses

Children exhibit unique activity patterns in classrooms, and their movements influence the resuspension of particulate matter, thereby increasing the risk of inhalation. To quantify children’s activity patterns and particulate matter inhalation doses, we measured particulate matter concentrations and tracked the recess activities of 194 children in two primary schools. YOLO v9 combined with the DeepSORT algorithm was used to identify and track the children in the videos, thereby calculating the speed and duration of each child. Statistical analysis revealed that classroom activities were highly transitory and predominantly of light intensity. The proportion of light-intensity physical activity in this study was 7%–15% higher than that reported in previous studies, attributable to high occupant density, limited activity space, and characteristics of the Chinese education environment. The median durations of recess activities decreased from light-intensity (4.7 s) to moderate-intensity (2.5 s) and vigorous-intensity activities (2.2 s). Furthermore, children’s activity speed and duration were strongly associated with variations in indoor PM2.5 concentrations. Additionally, the daily inhalation dose during recess decreased with increasing age, being 14.67% lower in the middle age group and 30.64% lower in the upper age group compared to the lower age group. Our analysis provides a valuable reference for assessing the health risks caused by particulate matter and for more effective measures to improve the classroom environment.