인문학
사회과학
자연과학
공학
의약학
농수해양학
예술체육학
복합학
지원사업
학술연구/단체지원/교육 등 연구자 활동을 지속하도록 DBpia가 지원하고 있어요.
커뮤니티
연구자들이 자신의 연구와 전문성을 널리 알리고, 새로운 협력의 기회를 만들 수 있는 네트워킹 공간이에요.
논문 기본 정보
- 자료유형
- 학술저널
- 저자정보
- 발행연도
- 2026.6
- 수록면
- 164 - 171 (8page)
- DOI
- 10.7315/CDE.2026.164
이용수
초록· 키워드
Analyzing worker trajectories in industrial sites is essential for productivity and safety. However, single-camera tracking suffers from trajectory fragmentation—where identification (ID) frequently changes—due to severe occlusions by complex equipment and intersecting fields of view. This study proposes a Global Trajectory Reconstruction Method to accurately merge fragmented tracklets from multiple cameras into a continuous global trajectory. Initial tracklets are extracted using BoT-SORT and projected onto a 2D global coordinate system via a Homography matrix. Disconnected trajectories are then optimally linked using spatio-temporal stitching based on a Kalman filter and Hungarian algorithm. For validation, a graphical assessment was conducted by overlaying reconstructed trajectories and the ground truth onto a 2D plot. The analysis confirmed the successful elimination of broken paths and parallel ghosting, allowing for the precise extraction of continuous global movements. This study is significant in accurately reconstructing reliable global trajectory data from fragmented vision data, providing foundational metrics for industrial site management and process analysis.
#Global trajectory reconstruction method
#Homography matrix
#Hungarian algorithm
#Kalman filter
#Multi-camera
#Object tracking
상세정보 수정요청해당 페이지 내 제목·저자·목차·페이지정보가 잘못된 경우 알려주세요!
목차
- ABSTRACT
- 1. 서론
- 2. 전역적 궤적 재구축 방법
- 3. 가상 공장 내 작업자 경로 추적
- 4. 결론 및 향후 계획
- References