Ruihao Li, Qiang Liu, Jianjun Gui, Dongbing Gu, Huosheng Hu
{"title":"Night-time indoor relocalization using depth image with Convolutional Neural Networks","authors":"Ruihao Li, Qiang Liu, Jianjun Gui, Dongbing Gu, Huosheng Hu","doi":"10.1109/ICONAC.2016.7604929","DOIUrl":null,"url":null,"abstract":"In this work, we present a Convolutional Neural Network(CNN) with depth images as its inputs to solve the relocalization problem of a moving platform in night-time indoor environment. The developed algorithm can estimate the camera pose in an end-to-end manner with 0.40m and 7.49° errors in real time during night. It does not require any geometric computation as it directly uses a CNN for 6 DOFs pose regression. The architecture and its encoding methods of depth images are discussed. The proposed method is also evaluated on benchmark datasets collected from a motion capture system in our lab.","PeriodicalId":375052,"journal":{"name":"2016 22nd International Conference on Automation and Computing (ICAC)","volume":"108 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2016-10-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"7","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2016 22nd International Conference on Automation and Computing (ICAC)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICONAC.2016.7604929","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 7
Abstract
In this work, we present a Convolutional Neural Network(CNN) with depth images as its inputs to solve the relocalization problem of a moving platform in night-time indoor environment. The developed algorithm can estimate the camera pose in an end-to-end manner with 0.40m and 7.49° errors in real time during night. It does not require any geometric computation as it directly uses a CNN for 6 DOFs pose regression. The architecture and its encoding methods of depth images are discussed. The proposed method is also evaluated on benchmark datasets collected from a motion capture system in our lab.