1. Inicio
  2. Atrás
Título: Detecting and locating moving objects in images taken from a mobile robot
Autores: Bhanji, Altaz.
Fecha: 1997
Publicador: McGill University - MCGILL
Fuente:
Tipo: Electronic Thesis or Dissertation
Tema: Engineering, Electronics and Electrical.
Computer Science.
Descripción: This thesis describes an algorithm for detecting moving objects in images taken from a moving robot. The algorithm distinguishes itself from other similar, existing solutions in that it lends itself to a fast implementation, yet provides results which are reliable and accurate enough for use on a real mobile robot system without any a priori knowledge of the robot's egomotion. The algorithm builds on previous work by Frazier and Nevatia but adds two key enhancements in order to realize a working system. First is the ability to efficiently and continuously update the robot's egomotion parameters by way of the focus of expansion (FOE). This is accomplished by minimizing an objective function which computes the percentage of "apparently moving" edges in the dynamic map for various estimates of the FOE position. The search space of this objective function is significantly reduced by utilizing the notion of the Fuzzy FOE. Second is the inclusion of a maximum likelihood (ML) classifier to provide a means for automatic thresholding to differentiate between static and moving objects. Results for both simulated and real data are presented and indicate that algorithm is fast and reliable.
Idioma: en