ETRI-Knowledge Sharing Plaform

ENGLISH

성과물

논문 검색
구분 SCI
연도 ~ 키워드

상세정보

학술대회 Bin Picking Method Using Multiple Local Features
Cited 1 time in scopus Download 0 time Share share facebook twitter linkedin kakaostory
저자
김계경, 강상승, 김재홍, 이재연, 김중배
발행일
201510
출처
International Conference on Ubiquitous Robots and Ambient Intelligence (URAI) 2015, pp.148-150
DOI
https://dx.doi.org/10.1109/URAI.2015.7358848
협약과제
15PC1600, 양팔 작업을 위한 센서융합 인지 기반 제어기술 개발 및 다중로봇 협업 생산공정 적용 기술 개발, 김중배
초록
Bin-picking technology using vision sensor for picking objects has studied intensively because it results in productivity improvement by applying automation process in industry fields. To obtain more accurate result of position detection and pose estimation of objects to be picked using robot system is not trivial task because of poor factors such as nonuniform lighting condition, occlusion, pose variation. In this paper, vision based object detection and pose estimation method for bin-picking are proposed that provides high accuracy for detecting object position and estimating distance be offered to industrial robot. Multiple local features are extracted and recognized for detecting object position and estimating pose of a picking object among randomly piled objects in a supply bin. We have simulated to evaluate performance of position detection and pose estimation of object using database captured under various lighting condition and in a pilot system, which has built alike a production site.
키워드
Bin-picking, Multiple local feature extraction, Object location detection, Pose estimation
KSP 제안 키워드
Bin-picking, Estimation method, High accuracy, Industrial robot, Lighting conditions, Multiple local feature extraction, Object detection, Object location detection, Pose estimation, Productivity improvement, Robot System