| Author: | Wu, Miao |
| Title: | Visual positioning based on sky-visibility matching with 3D city models |
| Degree: | M.Sc. |
| Year: | 2023 |
| Department: | Department of Mechanical Engineering |
| Pages: | vi, 57 pages : color illustrations |
| Language: | English |
| Abstract: | Localization and positioning are in great demand today, such as location-based services (LBS), emergency rescue and autonomous driving. Global navigation satellite system (GNSS) as an essential technology for positioning can reach meter-level accuracy in open sky environments. However, due to high-rise buildings and narrow streets, the positioning error increases dramatically, sometimes even reaching dozens of meters. This error is often caused by none-line-of-sight (NLOS) reception and multipath interference. Extra equipment and 3D city model are considered by providing additional information to improve the positioning accuracy. Here, we choose surrounding building boundaries as features to aid positioning. This paper proposes an approach for improving positioning accuracy by matching sky-visibility between the query photo and some particles’ Skymask. First, the building boundary is extracted from the taken photo and then plot a Skymask based on it. Meanwhile, some position candidates are distributed around the initial position guess obtained by Google Maps. Skymask for each particle can retrieve from precomputed Skymask database which is generated by a 3D city model. Then compare the elevation difference between image-converted Skymask with particles’ Skymask, namely sky-visibility matching, and weighting is given to each candidate. Resampling is done to leave those closer particles based on their weighting. Finally, calculate the weighted average of the remaining particles’ coordinates to determine the final position solution. Positioning experiments were conducted in urban environments in Hung Hom and Tsim Sha Tsui, Hong Kong to evaluate the algorithm's feasibility and the localization results' accuracy. We use a panoramic camera to take photos and some smartphones to collect positioning data. The Skymask database is generated from a 3D city model. Experimental results are then compared with conventional location solutions and NMEA solutions. It shows that the proposed method can achieve a positioning accuracy of fewer than 2 meters in most tested urban canyon areas, which is better than the other two methods. Moreover, the localization solution of the method in this paper is more stable. |
| Rights: | All rights reserved |
| Access: | restricted access |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 7481.pdf | For All Users (off-campus access for PolyU Staff & Students only) | 3.77 MB | Adobe PDF | View/Open |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/13013