Qayyum, Usman; Kim, Jonghyuk
Monocular visual SLAM approaches are mostly constrained in their performance due to general motion model and availability of true scale information. We proposed an approach which improves the motion prediction step of visual SLAM and results in better estimation of map scale. The approach utilizes the short term accuracy of inertial velocity with visual orientation to estimate refined motion priors. These motion priors are fused with sparse number of 3D map features to constraint the positional...[Show more]
Items in Open Research are protected by copyright, with all rights reserved, unless otherwise indicated.