Next Article in Journal
Deep-Learning-Based Physical Layer Authentication for Industrial Wireless Sensor Networks
Next Article in Special Issue
Assessment of BeiDou-3 and Multi-GNSS Precise Point Positioning Performance
Previous Article in Journal
Validity of Instrumented Insoles for Step Counting, Posture and Activity Recognition: A Systematic Review
Previous Article in Special Issue
High-Precision Indoor Visible Light Positioning Using Modified Momentum Back Propagation Neural Network with Sparse Training Point
Open AccessArticle

ConvNet and LSH-Based Visual Localization Using Localized Sequence Matching

1
Australian Centre for Field Robotics (ACFR), Department of Aerospace, Mechanical and Mechatronic Engineering (AMME), The University of Sydney, Sydney, NSW 2006, Australia
2
Connaissance et Intelligence Artificielle Distribuées (CIAD), University Bourgogne Franche-Comté, UTBM, F-90010 Belfort, France
*
Author to whom correspondence should be addressed.
The Main Experiments and Manuscript Were Finished in UTBM, While the Final Submission Was Conducted in ACFR. Part of the Experiment Results of This Paper Has been Published in Industrial Electronics Society, IECON 42nd Annual Conference of the IEEE.
Sensors 2019, 19(11), 2439; https://doi.org/10.3390/s19112439
Received: 20 April 2019 / Revised: 21 May 2019 / Accepted: 23 May 2019 / Published: 28 May 2019
(This article belongs to the Collection Positioning and Navigation)
Convolutional Network (ConvNet), with its strong image representation ability, has achieved significant progress in the computer vision and robotic fields. In this paper, we propose a visual localization approach based on place recognition that combines the powerful ConvNet features and localized image sequence matching. The image distance matrix is constructed based on the cosine distance of extracted ConvNet features, and then a sequence search technique is applied on this distance matrix for the final visual recognition. To speed up the computational efficiency, the locality sensitive hashing (LSH) method is applied to achieve real-time performances with minimal accuracy degradation. We present extensive experiments on four real world data sets to evaluate each of the specific challenges in visual recognition. A comprehensive performance comparison of different ConvNet layers (each defining a level of features) considering both appearance and illumination changes is conducted. Compared with the traditional approaches based on hand-crafted features and single image matching, the proposed method shows good performances even in the presence of appearance and illumination changes. View Full-Text
Keywords: visual localization; place recognition; convolutional network; sequence matching; LSH; SLAM visual localization; place recognition; convolutional network; sequence matching; LSH; SLAM
Show Figures

Figure 1

MDPI and ACS Style

Qiao, Y.; Cappelle, C.; Ruichek, Y.; Yang, T. ConvNet and LSH-Based Visual Localization Using Localized Sequence Matching. Sensors 2019, 19, 2439. https://doi.org/10.3390/s19112439

AMA Style

Qiao Y, Cappelle C, Ruichek Y, Yang T. ConvNet and LSH-Based Visual Localization Using Localized Sequence Matching. Sensors. 2019; 19(11):2439. https://doi.org/10.3390/s19112439

Chicago/Turabian Style

Qiao, Yongliang; Cappelle, Cindy; Ruichek, Yassine; Yang, Tao. 2019. "ConvNet and LSH-Based Visual Localization Using Localized Sequence Matching" Sensors 19, no. 11: 2439. https://doi.org/10.3390/s19112439

Find Other Styles
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.

Article Access Map by Country/Region

1
Search more from Scilit
 
Search
Back to TopTop