Analysis of the temporal correlations of TLS range observations from plane fitting residuals

authored by
Gaël Kermarrec, Michael Lösler, Jens Hartmann
Abstract

Terrestrial laser scanners (TLS) record a large number of points within a short time. Temporal correlations between observations are unavoidable but often neglected in stochastic modelling. The main consequences are an overestimated precision of the point clouds and potential wrong test decisions when used for deformation analysis with rigorous statistical procedures. Regarding physical considerations, a fractional Gaussian noise, defined by a so-called Hurst exponent, or a combination of fractional Gaussian noises could be used to model the noise of range measurements from a sensor perspective; Temporal correlations are expected to have a long-range dependency due to the high recording rate of the TLS. Scanning settings and configurations can affect the global correlation parameters. These effects can be quantified from the residuals of a least-squares surface approximation from the TLS point cloud. Based on simulation results, real data correlation analysis from indoor and outdoor experiments can be better understood which makes the identification of the dominant correlating noise source possible. Our methodology combines two Hurst-estimators: the Whittle maximum likelihood and the generalised Hurst estimator; It paves the way for a simple and global model for describing the temporal noise of TLS range correlations, usable in point clouds analysis independently of the object under consideration.

Organisation(s)
Geodetic Institute
QuantumFrontiers
External Organisation(s)
Frankfurt University of Applied Sciences
Type
Article
Journal
ISPRS Journal of Photogrammetry and Remote Sensing
Volume
171
Pages
119-132
No. of pages
14
ISSN
0924-2716
Publication date
01.2021
Publication status
Published
Peer reviewed
Yes
ASJC Scopus subject areas
Atomic and Molecular Physics, and Optics, Engineering (miscellaneous), Computer Science Applications, Computers in Earth Sciences
Electronic version(s)
https://doi.org/10.15488/11164 (Access: Open)
https://doi.org/10.1016/j.isprsjprs.2020.10.012 (Access: Closed)