School-age children, due to their physiological and psychological immaturity and relatively limited traffic experience, face higher risks when crossing signalized intersections. To gain a deeper understanding of the characteristics of school-age children's street-crossing behavior, this study focused on signalized intersections near elementary schools. Field video surveys were conducted to collect videos of children crossing the streets. The collected video data were processed using SIMI Motion software to extract parameters related to crossing behavior, including walking trajectories, crossing speed, crossing time, and horizontal and vertical displacement. Furthermore, combined with manual calibration methods, different types of crossing behavior, such as normal crossing and delayed crossing, were classified. Corresponding speed analysis charts were created using Origin. SPSS software was used to perform significance analysis on the differences in crossing behavior parameters among different types of crossings, quantifying the similarities and differences between them. The research results help to understand school-age childrens street-crossing behavior and provide targeted improvements for crossing facilities and traffic signal timing at signalized intersections around elementary schools.