Advanced search
Start date
Betweenand
Related content


ROAD NETWORK EXTRACTION USING GPS TRAJECTORIES BASED ON MORPHOLOGICAL AND SKELETONIZATION ALGORITHMS

Full text
Author(s):
Dal Pozi, A. P. ; Martins, E. F. O. ; Zanin, R. B. ; Ziatanova, S ; Sithole, G ; Barton, J
Total Authors: 6
Document type: Journal article
Source: XXIV ISPRS CONGRESS: IMAGING TODAY, FORESEEING TOMORROW, COMMISSION III; v. 43-B4, p. 7-pg., 2022-01-01.
Abstract

In this article, a method for road network extraction is proposed, based on GPS (Global Positioning System) trajectories. Unlike existing methods, it is not necessary to resample the GPS trajectories into a raster structure; instead, all analyses are based on the polylines that represent the GPS trajectories. Basically, a morphological analysis and a skeletonization technique are used by the proposed method. Two main steps of the method can be identified: the first step consists in generating an elongated polygon (that delimitates an elongated ribbon) that represents the selected road; and the second step aims at reconstructing the road network. The proposed method was evaluated based on four GPS trajectory datasets and the results obtained can be considered good, but some inconsistencies were noted, as for example: extraction failures occur in places with very low trajectory density (such as 3-4 trajectories); merging of very close and parallel roads; some road crossings that are close to one another have been merged into a single point. The proposed method was also compared with existing methods in the literature and the obtained results showed good consistency between them. (AU)

FAPESP's process: 21/03586-2 - Deep convolutional neural network (DCNN) for road network extraction from fusion of airborne laser scanning (ALS) data and highest resolution image in the urban environment
Grantee:Aluir Porfírio Dal Poz
Support Opportunities: Regular Research Grants