RT Conference Proceedings
T1 Accurate Stereo Visual Odometry with Gamma Distributions
A1 Gómez-Ojeda, Rubén
A1 Moreno-Dueñas, Francisco Ángel
A1 González-Jiménez, Antonio Javier
K1 Campo visual
AB Point-based stereo visual odometry systemstypically estimate the camera motion by minimizing a cost function of the projection residuals between consecutive frames. Under some mild assumptions, such minimization is equivalent to maximizingthe probability of the measured residuals givena certain pose change, for which a suitable model of the error distribution (sensor model) becomes of capital importance in order to obtain accurate results. This paper proposes a robust probabilistic model for projection errors, based on real world data. For that,we argue that projection distances follow Gammadistributions, and hence, the introduction of thesemodels in a probabilistic formulation of the motionestimation process increases both precision and accuracy. Our approach has been validated through a series of experiments with both synthetic and real data, revealing an improvement in accuracy while not increasing the computational burden.
PB IEEE
YR 2017
FD 2017
LK http://hdl.handle.net/10630/14376
UL http://hdl.handle.net/10630/14376
LA eng
NO Universidad de Málaga. Campus de Excelencia Internacional Andalucía Tech.               Project "PROMOVE: Advances in mobile robotics for promoting independent life of elders", funded by the Spanish Government and the "European RegionalDevelopment Fund ERDF" under contract DPI2014-55826-R.
DS RIUMA. Repositorio Institucional de la Universidad de Málaga
RD 20 ene 2026