| Title |
Improving the Design of Station Plazas as In-Between Spaces Through YOLO and Pose Estimation Analysis of Spatial Elements and User Behavior |
| Authors |
황미리(Miri, Hwang) ; 박은주(Eun Joo, Park) |
| DOI |
https://doi.org/10.5659/JAIK.2025.41.12.47 |
| Keywords |
Public Space; In-between Space; Station Plaza; User Behavior; Yolo Algorithm; Pose Estimation |
| Abstract |
This study examines how station plazas serve as transitional urban spaces where movement and stay behaviors intersect beyond simple transit
use. YOLO-based object detection and pose estimation techniques were applied to analyze pedestrian activity in three major plazas in Seoul:
Oullim Plaza at DDP Station, Star Plaza at Sinchon Station, and Culture Station Seoul 284 Plaza at Seoul Station. Pedestrian behavior was
categorized into four types: walking, running, standing, and sitting, to quantitatively assess how spatial elements such as walkways, benches,
and plantings influence user activity at different times of day, including morning, afternoon, and evening, as well as across weekdays and
weekends. The analysis showed that all three plazas were primarily movement-oriented, with activity concentrated along main pedestrian
routes rather than in areas designed for staying. Benches and planting zones were notably underused, highlighting a gap between spatial
design intent and actual user behavior. Based on these findings, the study suggests design strategies that include the targeted placement of
stay-promoting features, redistribution of spatial functions, and time-responsive spatial planning. By combining AI-based behavioral analysis
with urban design principles, it presents a framework for transforming station plazas into human-centered public spaces that encourage
pausing, interaction, and observation, rather than serving only as transit nodes. |