Multiple descriptors for Visual Odometry Trajectory Estimation

Research output: Contribution to journalArticle

Abstract

Visual Simultaneous Localization and Mapping (VSLAM) systems are widely used in mobile robots for autonomous navigation. One important part of VSLAM is trajectory estimation. Trajectory estimation is a part of the localization task in VSLAM where a robot needs to estimate the camera pose in order to precisely align the real visited image locations. The poses are estimated using Visual Odometer Trajectory Estimation (VOTE) by extracting distinctive and traceable key points from sequence image locations having been visited by a robot. In the visual trajectory estimation, one of the most popular solutions is arguably PnPRANSCA function. PnP-RANSAC is a common approach used for estimating the VOTE, which uses a feature descriptor such as SURF to extract key-points and match them in pairs based on their descriptors. However, due to the sensor noise and the high fluctuating scenes constitute an inevitable shortcoming that reduces the single visual descriptor performance in extracting the distinctive and traceable key points. Thus, this paper proposes a method that uses a random sampling scheme to combine the result of multiple key-points descriptors. The scheme extracts the best key points from SIFT, SURF and ORB key-point detectors based on their key-point response value. These key points are combined and refined based on Euclidean distances. This combination of key points with their corresponding visual descriptors is used in VOTE, which reduces the trajectory estimation errors. The proposed algorithm is evaluated on the widely used benchmark dataset KITTI where the three longest sequences are selected, 00 with 4541 images, 02 with 2761 images, and 05 with 1101 images. In trajectory estimation experiment, the proposed algorithm can reduce the trajectory error of 44%, 8% and 13% on KITTI dataset for the sequence 00, 02 and 05 respectively based on translational and rotational errors. In addition, the proposed algorithm succeeded in reducing the number of key points used in VOTE as combined with the state-of-the-art RTAB-Map.

Original languageEnglish
Pages (from-to)1423-1430
Number of pages8
JournalInternational Journal on Advanced Science, Engineering and Information Technology
Volume8
Issue number4-2
Publication statusPublished - 1 Jan 2018

Fingerprint

trajectories
Trajectories
Benchmarking
robots
Noise
Robots
extracts
Datasets
cameras
Mobile robots
Error analysis
detectors
sensors (equipment)
Navigation
Cameras
Sampling
Detectors
Sensors

Keywords

  • Feature matching
  • RANSAC
  • Selection scheme
  • Structure from motion
  • Trajectory estimation
  • Visual Odometer

ASJC Scopus subject areas

  • Computer Science(all)
  • Agricultural and Biological Sciences(all)
  • Engineering(all)

Cite this

Multiple descriptors for Visual Odometry Trajectory Estimation. / Salameh, Mohammed; Abdullah, Azizi; Sahran, Shahnorbanun.

In: International Journal on Advanced Science, Engineering and Information Technology, Vol. 8, No. 4-2, 01.01.2018, p. 1423-1430.

Research output: Contribution to journalArticle

@article{52d8dbe8473e494e9c780da8cfc08708,
title = "Multiple descriptors for Visual Odometry Trajectory Estimation",
abstract = "Visual Simultaneous Localization and Mapping (VSLAM) systems are widely used in mobile robots for autonomous navigation. One important part of VSLAM is trajectory estimation. Trajectory estimation is a part of the localization task in VSLAM where a robot needs to estimate the camera pose in order to precisely align the real visited image locations. The poses are estimated using Visual Odometer Trajectory Estimation (VOTE) by extracting distinctive and traceable key points from sequence image locations having been visited by a robot. In the visual trajectory estimation, one of the most popular solutions is arguably PnPRANSCA function. PnP-RANSAC is a common approach used for estimating the VOTE, which uses a feature descriptor such as SURF to extract key-points and match them in pairs based on their descriptors. However, due to the sensor noise and the high fluctuating scenes constitute an inevitable shortcoming that reduces the single visual descriptor performance in extracting the distinctive and traceable key points. Thus, this paper proposes a method that uses a random sampling scheme to combine the result of multiple key-points descriptors. The scheme extracts the best key points from SIFT, SURF and ORB key-point detectors based on their key-point response value. These key points are combined and refined based on Euclidean distances. This combination of key points with their corresponding visual descriptors is used in VOTE, which reduces the trajectory estimation errors. The proposed algorithm is evaluated on the widely used benchmark dataset KITTI where the three longest sequences are selected, 00 with 4541 images, 02 with 2761 images, and 05 with 1101 images. In trajectory estimation experiment, the proposed algorithm can reduce the trajectory error of 44{\%}, 8{\%} and 13{\%} on KITTI dataset for the sequence 00, 02 and 05 respectively based on translational and rotational errors. In addition, the proposed algorithm succeeded in reducing the number of key points used in VOTE as combined with the state-of-the-art RTAB-Map.",
keywords = "Feature matching, RANSAC, Selection scheme, Structure from motion, Trajectory estimation, Visual Odometer",
author = "Mohammed Salameh and Azizi Abdullah and Shahnorbanun Sahran",
year = "2018",
month = "1",
day = "1",
language = "English",
volume = "8",
pages = "1423--1430",
journal = "International Journal on Advanced Science, Engineering and Information Technology",
issn = "2088-5334",
publisher = "INSIGHT - Indonesian Society for Knowledge and Human Development",
number = "4-2",

}

TY - JOUR

T1 - Multiple descriptors for Visual Odometry Trajectory Estimation

AU - Salameh, Mohammed

AU - Abdullah, Azizi

AU - Sahran, Shahnorbanun

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Visual Simultaneous Localization and Mapping (VSLAM) systems are widely used in mobile robots for autonomous navigation. One important part of VSLAM is trajectory estimation. Trajectory estimation is a part of the localization task in VSLAM where a robot needs to estimate the camera pose in order to precisely align the real visited image locations. The poses are estimated using Visual Odometer Trajectory Estimation (VOTE) by extracting distinctive and traceable key points from sequence image locations having been visited by a robot. In the visual trajectory estimation, one of the most popular solutions is arguably PnPRANSCA function. PnP-RANSAC is a common approach used for estimating the VOTE, which uses a feature descriptor such as SURF to extract key-points and match them in pairs based on their descriptors. However, due to the sensor noise and the high fluctuating scenes constitute an inevitable shortcoming that reduces the single visual descriptor performance in extracting the distinctive and traceable key points. Thus, this paper proposes a method that uses a random sampling scheme to combine the result of multiple key-points descriptors. The scheme extracts the best key points from SIFT, SURF and ORB key-point detectors based on their key-point response value. These key points are combined and refined based on Euclidean distances. This combination of key points with their corresponding visual descriptors is used in VOTE, which reduces the trajectory estimation errors. The proposed algorithm is evaluated on the widely used benchmark dataset KITTI where the three longest sequences are selected, 00 with 4541 images, 02 with 2761 images, and 05 with 1101 images. In trajectory estimation experiment, the proposed algorithm can reduce the trajectory error of 44%, 8% and 13% on KITTI dataset for the sequence 00, 02 and 05 respectively based on translational and rotational errors. In addition, the proposed algorithm succeeded in reducing the number of key points used in VOTE as combined with the state-of-the-art RTAB-Map.

AB - Visual Simultaneous Localization and Mapping (VSLAM) systems are widely used in mobile robots for autonomous navigation. One important part of VSLAM is trajectory estimation. Trajectory estimation is a part of the localization task in VSLAM where a robot needs to estimate the camera pose in order to precisely align the real visited image locations. The poses are estimated using Visual Odometer Trajectory Estimation (VOTE) by extracting distinctive and traceable key points from sequence image locations having been visited by a robot. In the visual trajectory estimation, one of the most popular solutions is arguably PnPRANSCA function. PnP-RANSAC is a common approach used for estimating the VOTE, which uses a feature descriptor such as SURF to extract key-points and match them in pairs based on their descriptors. However, due to the sensor noise and the high fluctuating scenes constitute an inevitable shortcoming that reduces the single visual descriptor performance in extracting the distinctive and traceable key points. Thus, this paper proposes a method that uses a random sampling scheme to combine the result of multiple key-points descriptors. The scheme extracts the best key points from SIFT, SURF and ORB key-point detectors based on their key-point response value. These key points are combined and refined based on Euclidean distances. This combination of key points with their corresponding visual descriptors is used in VOTE, which reduces the trajectory estimation errors. The proposed algorithm is evaluated on the widely used benchmark dataset KITTI where the three longest sequences are selected, 00 with 4541 images, 02 with 2761 images, and 05 with 1101 images. In trajectory estimation experiment, the proposed algorithm can reduce the trajectory error of 44%, 8% and 13% on KITTI dataset for the sequence 00, 02 and 05 respectively based on translational and rotational errors. In addition, the proposed algorithm succeeded in reducing the number of key points used in VOTE as combined with the state-of-the-art RTAB-Map.

KW - Feature matching

KW - RANSAC

KW - Selection scheme

KW - Structure from motion

KW - Trajectory estimation

KW - Visual Odometer

UR - http://www.scopus.com/inward/record.url?scp=85055350951&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85055350951&partnerID=8YFLogxK

M3 - Article

VL - 8

SP - 1423

EP - 1430

JO - International Journal on Advanced Science, Engineering and Information Technology

JF - International Journal on Advanced Science, Engineering and Information Technology

SN - 2088-5334

IS - 4-2

ER -