Eventos Anais de eventos
COBEM 2017
24th ABCM International Congress of Mechanical Engineering
PERFORMANCE EVALUATION OF A MULTISCALE DESCRIPTOR TO DETECT FEATURE POINTS ON OBJECT BORDERS IN RANGE IMAGES
Submission Author:
José Maurício Santos Torres da Motta , DF , Brazil
Co-Authors:
Gerardo Idrobo Pizo, José Maurício Santos Torres da Motta
Presenter: Gerardo Idrobo Pizo
doi://10.26678/ABCM.COBEM2017.COB17-1193
Abstract
In this article, it is presented a new method to detect object borders in a multiscale level using descriptors that encode the neighborhood of points belonging to the borders of the object with numeric values named signature. The process generates as output feature points that are used for the registration of the range images. The proposal is to contribute to the state of the art of 3D reconstruction of workspaces that are mapped for navigation of service robots. For evaluating the performance of the proposed descriptor to detect borders in noisy images, the results were compared to the results obtained by other authors and techniques by using test images. The results are presented statistically as the number of feature points identified as true and false by each technique. It can be observed that the proposed technique provides more true points and less false points, approaching the results obtained with a manual identification.
Keywords
Object Borders, Multiscale Descriptor, Range Image Registration
