東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

A novel update scheme for landmark i...

Chen, Zhenhe.

Linked to FindBook

Google Book

Amazon

博客來

A novel update scheme for landmark initialization in monocular visual SLAM.

Record Type:	Language materials, printed : Monograph/item
Title/Author:	A novel update scheme for landmark initialization in monocular visual SLAM./
Author:	Chen, Zhenhe.
Description:	146 p.
Notes:	Source: Dissertation Abstracts International, Volume: 68-09, Section: B, page: 6174.
Contained By:	Dissertation Abstracts International68-09B.
Subject:	Engineering, Electronics and Electrical. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=NR30795
ISBN:	9780494307953

A novel update scheme for landmark initialization in monocular visual SLAM.
Chen, Zhenhe.

A novel update scheme for landmark initialization in monocular visual SLAM. - 146 p.

Source: Dissertation Abstracts International, Volume: 68-09, Section: B, page: 6174.

Thesis (Ph.D.)--The University of Western Ontario (Canada), 2007.

Through our theoretical investigation and experiments, we show that robotic SLAM using a monocular vision system is feasible and promising.

ISBN: 9780494307953Subjects--Topical Terms:

626636
Engineering, Electronics and Electrical.

A novel update scheme for landmark initialization in monocular visual SLAM.
LDR:03198nam 2200289 a 45 001 955018
005 20110622
008 110622s2007 ||||||||||||||||| ||eng d
020 $a 9780494307953
035 $a (UMI)AAINR30795
035 $a AAINR30795
040 $a UMI $c UMI
100 1 $a Chen, Zhenhe. $3 1278480
245 1 2 $a A novel update scheme for landmark initialization in monocular visual SLAM.
300 $a 146 p.
500 $a Source: Dissertation Abstracts International, Volume: 68-09, Section: B, page: 6174.
502 $a Thesis (Ph.D.)--The University of Western Ontario (Canada), 2007.
520 $a Through our theoretical investigation and experiments, we show that robotic SLAM using a monocular vision system is feasible and promising.
520 $a Making a robot navigate autonomously is a grand challenge in the robotic and computer vision research community. We approach the autonomous robot navigation problem within the simultaneous localization and map-building (SLAM) paradigm. The SLAM problem examines the ability of an autonomous robot starting in an unknown environment to incrementally build an environment map and simultaneously localize itself within this map. In robot navigation, accurate mapping and location estimates in either structured or unstructured environments is an essential requirement for high level tasks such as patrolling, investigating, delivering and fetching by an autonomous robot.
520 $a This thesis explores the field of computer vision based robot navigation. It presents not only theoretical investigation but also practical implementation of Kalman filter based visual SLAM algorithms. The work presented in this thesis spans several areas of engineering and computer science, from new computer vision algorithms to novel ideas of uncertainty analysis in robotic SLAM.
520 $a The existing SLAM algorithms heavily rely on sophisticated sensing systems, such as the fusion of global positioning system, inertial sensor, and laser detection and ranging. We propose to solve the SLAM problem using monocular vision. We will show four principal contributions that allow us to use such a simple sensing system. The first is to extend the existing theoretical lower bound of map-building error by estimating a new numerical lower bound with fewer constraints. Our second contribution is to apply multiple view geometry (MVG) techniques resulting in a novel update scheme for landmark initialization. We use theoretical derivation to come up with a closed form solution for the error at each observation step and use numerical verifications to prove that the new scheme is able to reduce mapping error significantly. Third, we incorporate high level geometric information such as common geometric primitives to SLAM to enhance the robustness. Finally, the proposed landmark initialization algorithm is demonstrated and evaluated by implementing the full SLAM algorithm in an indoor environment using a monocular vision sensor.
590 $a School code: 0784.
650 4 $a Engineering, Electronics and Electrical. $3 626636
650 4 $a Engineering, Robotics. $3 1018454
690 $a 0544
690 $a 0771
710 2 $a The University of Western Ontario (Canada). $3 1017622
773 0 $t Dissertation Abstracts International $g 68-09B.
790 $a 0784
791 $a Ph.D.
792 $a 2007
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=NR30795