See-Your-Room:Indoor Localization with Camera Vision

摘要：

　　Indoor Localization has attracted much research effort recently.Most existing approaches require pre-deployed comprehensive lo-cation services and,more importantly,rely on dedicatedly installed infrastructure.This incurs a high deployment,maintenance and usage overhead.We present See-Your-Room,a precise and robust localization approach with public cameras which are already ex-isted in today's buildings.Computer Vision systems,however,are difficult to achieve people's position directly from photos.To detect people in the room and achieve their poses efficiently,we combine detection method with pose estimation method.To transform peo-ple's position in the photos which are taken by public camera to the indoor map,we use projection transformation with four pairs of points.We implement See-Your-Room on commercial cameras,say ceiling cameras in three diverse rooms.The evaluation results show that See-Your-Room achieves an immediate detection suc-cess rate of 90%and an average localization error of 14.32cm in 3D environment.

关键词： Camera Positioning Indoor Localization Computer Vision

作者: Min Sun Li Zhang Ying Liu Xin Miao Xuan Ding

作者单位: School of Software,Tsinghua University HeFei University of Technology Liaoning Technical University

会议类型: 国际会议

会议名称: 2019国图灵大会(ACM Turing Celebration conference-China 2019 )

会议地点: 成都

会议语种:英文

页码: 439-443

在线出版日期: 2019-05-17（万方平台首次上网日期，不代表论文的发表时间）