Evaluation of UAV-borne photogrammetry and UAV-borne laser scanning for 3D topographic change analysis of an active rock glacier

Recent advances in repeated data acquisition by UAV-borne photogrammetry and laser scanning for geoscientific monitoring extend the possibilities for analysing surface dynamics in 3D at high spatial (centimeter point spacing) and temporal (up to daily) resolution. These techniques overcome common challenges of ground-based sensing (occlusion, heterogeneous measurement distribution, limited spatial coverage) and provide a valuable additional data source for topographic change analysis between successive epochs.

We investigate point clouds derived from UAV-borne photogrammetry and laser scanning as input for change analysis. We apply and compare two state-of-the-art methods for pairwise 3D topographic change quantification. Our study site is the active rock glacier Äußeres Hochebenkar in the Eastern Austrian Alps (46° 50’ N, 11° 01’ E). Whereas point clouds derived from terrestrial laser scanning (TLS) have become a common data source for this application, point clouds derived from UAV-borne sensing techniques have emerged only in recent years and their potential for methods of 3D and 4D (3D + time) change analysis is yet to be exploited.

We perform change analysis using (1) the Multi Scale Model to Model Cloud Comparison (M3C2) algorithm [1] and (2) the correspondence-driven plane-based M3C2 [2]. Both methods have shown to provide valuable surface change information on rock glaciers when applied to successive terrestrial laser scanning point clouds of different time spans (ranging from 2 weeks to several years). The considerable value of both methods also lies in their ability to quantify the uncertainty additionally to the associated change. This allows to distinguish between significant change (quantified magnitude of change > uncertainty) and non-significant or no change (magnitude of change ≤ uncertainty) and hence enables confident analysis and geographic interpretation of change.

We will extend the application of the two methods by using point clouds derived using (1) photogrammetric techniques on UAV-based images and (2) UAV-borne laser scanning. We investigate the influence of variations in measurement distribution and density, completeness of spatial coverage and ranging uncertainty by comparing UAV-based point clouds to TLS data of the same epoch. Using TLS-TLS-based change analysis as reference, we examine the performance of the two methods with respect to their capability of quantifying surface change based on point clouds originating from different sensing techniques.

Results of this assessment can support the theoretical and practical design of future measurement set-ups. Comparing results of both methods further aids the selection of a suitable method (or combination) for change analysis in order to meet requirements e.g., regarding uncertainty of measured change or spatial coverage of the analysis. To ease usability of a broad suite of state-of-the-art methods of 3D/4D change analysis, we are implementing an open source Python library for geographic change analysis in 4D point cloud data (py4dgeo, www.uni-heidelberg.de/3dgeo-opensource). Finally, our presented study provides insights how methods for 3D and 4D change analysis should be adapted or developed in order to exploit the full potential of available close-range sensing techniques.

[1] https://doi.org/ 10.1016/j.isprsjprs.2013.04.009

[2] https://doi.org/10.1016/j.isprsjprs.2021.11.018