[Back]

C. Harmening, S. Kauker, H.-B. Neuner, V. Schwieger:
"Terrestrial Laserscanning - Modeling of Correlations and Surface Deformations";
Talk: 78th FIG Working Week 2016, Christchurch, Neuseeland; 2016-05-02 - 2016-05-06; in: "Proceedings of 78th FIG Working Week 2016", (2016), Paper ID 8030, 13 pages.

Terrestrial Laserscanning offers new possibilities to engineering geodesy in general and defor-mation analysis in particular. Huge amounts of measured points lead to changing modelling and analysis approaches. Within the project "Integrated spatio-temporal modelling using correlated ob-servations for the derivation of surveying configurations and description of deformation processes" (IMKAD) modeling of correlations and surfaces will be treated among others.
The modelling of correlations within laser scanning point clouds can be realized by using synthetic covariance matrices. These are based on the elementary error model that consists of non-correlating, functional correlating and stochastic correlating error groups. This elementary error model will be applied on terrestrial laser scanning by defining three groups of error sources: instrumental, atmos-pheric and object based ones. All known TLS-errors have to be classified and modelled according to the model of elementary errors. This contribution presents first simulation results for the Leica HDS 7000 measuring on small test pieces made of gypsum, aluminum and rusty steel. The deter-mined variances and the spatial correlations of the points are estimated and discussed. Hereby, the mean standard deviation of an individual point within the point cloud is up to 2.5 mm and the mean correlation is about 0.94 neglecting the object based error sources in a first approach.
In the second part of this paper the development of the trend component of a spatiotemporal contin-uous collocation in order to describe areal deformations is presented. This component is modelled by estimated B-spline surfaces. One set of parameters used to define B-Spline curves and surfaces are the control points. Their number and position need to be estimated from the measurements. Here the determination of the optimal number of control points is regarded as a model selection problem. Two linear model selection criteria - the Akaike Information Criterion (AIC) and the Bayesian In-formation Criterion (BIC) - are investigated, compared and applied to simulated data sets. Addi-tionally, the contribution will give an outlook with respect to the combination of the before men-tioned stochastic and the deterministic approaches with the aim to detect surface deformations in a stochastically correct way.

Deformation measurement, Laser scanning, Positioning