Recognising, Representing and Mapping in Field Robotics

Name: Recognising, Representing and Mapping in Field Robotics
Brand: VDM Verlag Dr. Müller
SKU: 9783639137590
Price: 81.57 EUR
Availability: InStock

A Statistical View to Perception in Unstructured Environments

Автор: Fabio Ramos

Език:

Корици: С меки корици

Издател: VDM Verlag Dr. Müller

Наличност: Налично при издателя, по поръчка

Изпращаме след 17-27 дни

81.57 € 159.54 лв

The problem of building statistical models for§multi-sensor §perception in unstructured outdoor envi...

Информация за книгата

Автор

Fabio Ramos

Език

Английски език

Корици

Книга - С меки корици

Издадена

2009

страници

212

EAN

9783639137590

Enbook ID

06822420

Издател

VDM Verlag Dr. Müller

Теглоt

332

Размери

150 x 220 x 13

Пълно описание

The problem of building statistical models for§multi-sensor §perception in unstructured outdoor environments is§addressed in §this book. The perception problem is divided into§three distinct §tasks: recognition, representation and association.§Recognition is §cast as a statistical classification problem where§inputs are images §or a combination of images and ranging information.§Given the §complexity and variability of natural environments,§the use of §Bayesian statistics and supervised dimensionality§reduction to §incorporate prior information and to fuse sensory§data are §investigated. This book presents techniques for§combining non-§linear dimensionality reduction with parametric§learning through §Expectation Maximisation to build general and compact §representations of natural features. The robustness§of localisation §and mapping algorithms is directly related to§reliable data §association. A new data association algorithm§incorporating visual §and geometric information is proposed to improve the§reliability of §this task. The method uses a compact probabilistic§representation of §objects to fuse visual and geometric information for§the association §decision.

Може също да ви хареса