Multiple straight-line fitting using a Bayes factor
Por:
Lara-Alvarez C., Romero L., Gomez C.
Publicada:
1 ene 2017
Resumen:
This paper introduces a Bayesian approach to solve the problem of fitting multiple straight lines to a set of 2D points. Other approaches use many arbitrary parameters and threshold values, the proposed criterion uses only the parameters of the measurement errors. Models with multiple lines are useful in many applications, this paper analyzes the performance of the new approach to solve a classical problem in robotics: finding a map of lines from laser measurements. Tests show that the Bayesian approach obtains reliable models. © 2016, Springer-Verlag Berlin Heidelberg.
Filiaciones:
CONACYT Research Fellow–Centro de Investigación en Matemáticas, A. C. (CIMAT), Av. Universidad 222, La Loma, Zacatecas, 96068, Mexico
Universidad Michoacana de San Nicolás de Hidalgo (UMSNH), Av. Francisco J. Múgica S/N, Ciudad Universitaria, Morelia, Mich 58030, Mexico
|