Title :
Improved visual odometry method for matching 3D point cloud data
Author :
Liang Zhiwei ; Xu Xiaogen
Author_Institution :
Coll. of Autom., Nanjing Univ. of Posts & Telecommun., Nanjing, China
Abstract :
Aiming at the problem of 3D map data association, this paper proposes a method based on improved visual odometry algorithm. Three-dimensional model of visual information is associated with six degrees of freedom of visual odometry to represent the transformation matrix. Using lie algebra coordinates of rigid body motion and maximizing photo consistency, the registration results are evaluated by linearization of the energy equation. This algorithm can improve the success rate of data fusion so that the 3D map can be built continuously and effectively in real time.
Keywords :
Lie algebras; image fusion; image matching; image registration; matrix algebra; robot vision; 3D map data association; 3D point cloud data matching; data fusion; energy equation linearization; photo consistency maximization; registration; rigid body motion Lie algebra coordinates; robot vision; transformation matrix; visual information three-dimensional model; visual odometry method; visual odometry six degrees of freedom; Cameras; Equations; Mathematical model; Robot sensing systems; Three-dimensional displays; Visualization; Data association; Loop closure detection; Natural characteristic; Visual odometry;
Conference_Titel :
Control Conference (CCC), 2014 33rd Chinese
Conference_Location :
Nanjing
DOI :
10.1109/ChiCC.2014.6896422
