Title :
Data-driven MCMC sampling for vision-based 6D SLAM
Author :
Min, Jeeeun ; Kim, Jung-Ho ; Shin, Seung Heon ; Kweon, In-So
Author_Institution :
Dept. of Electr. Eng., KAIST, Daejeon, South Korea
Abstract :
An efficient sampling technique for estimating camera motion is presented. For this purpose, Markov chain Monte Carlo (MCMC) sampling is incorporated into the data-driven proposal distribution in order to improve the SLAM performance. Experimental results using both synthetic and real datasets demonstrate the efficiency of the proposed method.
Keywords :
Markov processes; Monte Carlo methods; SLAM (robots); cameras; motion estimation; robot vision; Markov chain Monte Carlo sampling; camera motion estimation; data-driven MCMC sampling technique; data-driven proposal distribution; vision-based 6D SLAM;
Journal_Title :
Electronics Letters
DOI :
10.1049/el.2012.0897
