DocumentCode :
115834
Title :
Randomized algorithm for estimation of moving point position using single camera
Author :
Krivokon, Dmitry ; Vakhitov, Alexander
Author_Institution :
Fac. of Math. & Mech., St. Petersburg State Univ., St. Petersburg, Russia
fYear :
2014
fDate :
15-17 Dec. 2014
Firstpage :
5189
Lastpage :
5194
Abstract :
Stochastic approximation algorithms (for example SPSA) provide a way to solve optimization problems in the presence of arbitrary but bounded disturbances. In this paper a problem of position estimation for a moving point using monocular projective observations is considered. We add random perturbations to camera position to produce an algorithm which makes estimates of point position demanding only that the point´s velocity is bounded in time. This is superior to the methods currently available in the computer vision field which all consider very restricted cases of point movement (constant, movement in plane). We prove theoretical convergence of estimates and provide numerical simulation for the algorithm.
Keywords :
computer vision; image sensors; motion estimation; optimisation; randomised algorithms; arbitrary disturbances; bounded disturbances; computer vision field; monocular projective observations; moving point position estimation; optimization problems; point movement; random perturbations; randomized algorithm; single camera; stochastic approximation algorithms; Approximation algorithms; Cameras; Convergence; Estimation; Heuristic algorithms; Noise; Zinc;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Decision and Control (CDC), 2014 IEEE 53rd Annual Conference on
Conference_Location :
Los Angeles, CA
Print_ISBN :
978-1-4799-7746-8
Type :
conf
DOI :
10.1109/CDC.2014.7040200
Filename :
7040200
Link To Document :
بازگشت