This repository contains the source code an instruction how to use the tool for uncertainty estimation of displacement estimates by sampling local averages displacements. It accompanies the recent (submitted) publication:
Sensitivity Quantification of Spatially Averaged Displacement Estimates in TLS-based Geomonitoring (submitted)
Lorenz Schmid, Tomislav Medic, Nicholas Meyer, Andreas Wieser
Terrestrial Laser Scanning (TLS) is increasingly used in geomonitoring for 3D displacement analysis. However, assessing the sensitivity of the implemented monitoring strategy, which is crucial for correctly interpreting observations and detecting deformations, is challenging due to the impact of complex spatial correlations and other influencing factors. Traditional methods for sensitivity analysis often assume uncorrelated measurements, leading to biased and overestimated sensitivity. This can lead to suboptimal choices of monitoring strategy, false expectations, and errors in displacement detection. This study introduces a new method for quantifying the uncertainty of spatially aggregated (averaged) TLS-based displacement estimates in geomonitoring by empirically locally sampling such aggregated values (ELSA). The method implicitly accounts for the mentioned spatial correlations and their local variations, therefore, providing a more realistic sensitivity quantification. Validation using real-world datasets and simulated displacements demonstrates the method's ability to provide a realistic uncertainty estimate and, subsequently, a good sensitivity estimate, realized herein by means of the Minimal Detectable Bias (MDB). Finally, we investigated several data preprocessing steps and demonstrated their effectiveness in enhancing both sensitivity and the quality of uncertainty estimates.
Create a new environment
conda create --name <env name> python=3.11
conda activate <env name>git clone git@github.com:gseg-ethz/tls_sensitivity_processor.git
cd sensitivity_processor
pip install .Next, follow the instructions on how to run the system by typing:
sensitivity-processor --helpThis should print the following help message:
