Premium
Minimizing B‐spline knots in representative road axis from GPS points cloud
Author(s) -
RomeroZaliz R.,
Reinoso J. F.,
Barrera D.,
ArizaLópez F. J.
Publication year - 2015
Publication title -
mathematical methods in the applied sciences
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.719
H-Index - 65
eISSN - 1099-1476
pISSN - 0170-4214
DOI - 10.1002/mma.3772
Subject(s) - global positioning system , b spline , computer science , photogrammetry , spline (mechanical) , digital mapping , algorithm , mathematics , computer vision , data mining , geography , remote sensing , engineering , mathematical analysis , telecommunications , structural engineering
Accuracy in roads geometry is an objective to be achieved by surveyors and cartographers when they obtain their data by GPS, Lidar, or photogrammetry. Nevertheless, those capture methods are expensive. Nowadays, cheap and collaborative methods can produce big datasets, which need to be processed in order to get accuracy axis from not accurate original data. Because a roads network is composed of several points, the resulting dataset could become a large‐sized file, difficult to manage, and slow in consultancy for the users. In this paper, we expose our previous solutions for estimating a representative axis and propose a novel B‐spline least square method governed by a genetic algorithm. The genetic algorithm minimizes the number of knots necessary to define the B‐spline representative axis while keeping the axis' original shape. We know the original shape because we have computed it using a large number of knots by an iterative and convergent method developed in a well‐contrasted previous study. This paper shows that our approaches are suitable to be deployed in a web‐based application in order to support collaborative digital cartography. Copyright © 2015 John Wiley & Sons, Ltd.