ETRI-Knowledge Sharing Plaform

ENGLISH

성과물

논문 검색
구분 SCI
연도 ~ 키워드

상세정보

학술대회 Real Space 3D Mesh Transformation Preprocessing System for AR Services
Cited 1 time in scopus Download 5 time Share share facebook twitter linkedin kakaostory
저자
윤영석, 박상헌, 황상원, 이상윤, 조현우, 박찬호, 정성욱
발행일
202010
출처
International Conference on Information and Communication Technology Convergence (ICTC) 2020, pp.1478-1483
DOI
https://dx.doi.org/10.1109/ICTC49870.2020.9289462
협약과제
20IH2800, 사용자 참여형 문화공간 콘텐츠를 위한 AR 플랫폼 기술개발 , 정성욱
초록
We propose a system that preprocesses a real space into a 3D mesh to provide augmented reality service. The proposed system acquires an image sequence using an RGB camera to model a real space. First, we transform the obtained sequence of RGB images into a 3D point cloud with real spatial information. Also, we segment the objects in the RGB image sequence and classify classes of the objects. Next, the proposed system performs a three-dimensional point labeling, which allocates object information for each point by using the three-dimensional point cloud and the segmented information of the object. Then, points in 3D space with the same label are gathered for each object. Finally, 3D meshes are generated only from points having the same identification number among the collected points. In this paper, it was confirmed that the 3D mesh can be created only with not using various types of sensors to obtain 3D information but the RGB image sequence.
키워드
3D Mesh, 3D Point Labeling, Augmented Reality(AR) Service, Object Segmentation
KSP 제안 키워드
3D Mesh, 3D information, 3D point cloud, 3D space, Augmented reality(AR), Augmented reality service, Image sequence, Object segmentation, Point labeling, RGB camera, RGB image