Evaluating the Robustness of New Holistic Description Methods in Position Estimation
of Mobile Robots
V. Román, L. Payá, S. Cebollada, A. Peidró, O. Reinoso
Lecture Notes in Electrical Engineering. Informatics in Control, Automation and Robotics (2022)
Ed. Springer ISBN:978-3-030-92441-6 ISSN:1876-1100 DOI:https://doi.org/10.1007/978-3-030-92442-3_12 - 793 (207-225)
In this work, some holistic description methods are evaluated in the framework of a localization task in heterogeneous zones, task that an autonomous robot should be able to perform correctly. The unique source of information is an omnidirectional vision sensor and the work is focused on the use of holistic or global-appearance techniques to describe the visual information. Holistic descriptors consist in obtaining a unique vector that describes globally the image. The goal of the experiments is to check new approaches to build and to handle global descriptors. Previously, the holistic descriptors have been processed without considering the spatial distribution of the information. In contrast, in this work two different new methods are proposed which take the assumption that the most relevant information is on the central rows of the panoramic image. For this reason, in the proposed description methods, the central rows have a higher weight comparing to other zones of the image. The new techniques are compared with the classical method. The experiments are carried out in real environments, with sets of images captured while the robot traversed in different heterogeneous routes. Also, variations of the lighting conditions, people who occlude the scene and changes on the furniture may appear.