An approach for automatic characterization of surface activities from large 4D point clouds is presented in a new paper by Daan Hulskemper et al. in collaboration between the 3DGeo research group and the departments of Geoscience and Remote Sensing and Coastal Engineering at TU Delft.
Hulskemper, D., Anders, K., Antolínez, J. A. Á., Kuschnerus, M., Höfle, B., & Lindenbergh, R. (2022). Characterization of Morphological Surface Activities derived from Near-Continuous Terrestrial LiDAR Time Series. Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLVIII-2/W2-2022, pp. 53-60. doi: 10.5194/isprs-archives-XLVIII-2-W2-2022-53-2022.
The Earth’s landscapes are shaped by processes eroding, transporting and depositing material over various timespans and spatial scales. To understand these surface activities and mitigate potential hazards they inflict (e.g., the landward movement of a shoreline), knowledge is needed on the occurrences and impact of these activities. Near-continuous terrestrial laser scanning enables the acquisition of large datasets of surface morphology, represented as three-dimensional point cloud time series. Exploiting the full potential of this large amount of data, by extracting and characterizing different types of surface activities, is challenging. In this research we use a time series of 2,942 point clouds obtained over a sandy beach in The Netherlands (Vos et al., 2022). We investigate automated methods to extract individual surface activities present in this dataset and cluster them into groups to characterize different types of surface activities (e.g., intertidal sandbar deposition, anthropogenic beach nourishments). We show that, first extracting 2,021 spatiotemporal segments of surface activity as 4D objects-by-change (4D-OBCs; Anders et al., 2021), and second, clustering these segments with a Self-organizing Map (SOM) in combination with hierarchical clustering, allows for the unsupervised identification and characterization of different types of surface activities present on a sandy beach. The SOM enables us to find events displaying certain type of surface activity, while it also enables the identification of subtle differences between different events belonging to one specific surface activity. Hierarchical clustering then allows us to find and characterize broader groups of surface activity, even if the same type of activity occurs at different points in space or time.
The work will be presented at the Optical 3D Metrology (O3DM) Workshop in Würzburg on 15 December 2022. We are looking forward to meeting some of you there!
If you are interested in related work on monitoring of surface activity on a sandy beach, check out:
Anders, K., Winiwarter, L., Mara, H., Lindenbergh, R., Vos, S. E., & Höfle, B. (2021). Fully automatic spatiotemporal segmentation of 3D LiDAR time series for the extraction of natural surface changes. ISPRS Journal of Photogrammetry and Remote Sensing, 173, pp. 297-308. doi: 10.1016/j.isprsjprs.2021.01.015.
Vos, S., Anders, K., Kuschnerus, M., Lindenbergh, R., Höfle, B., Aarninkhof, S., & de Vries, S. (2022). A high-resolution 4D terrestrial laser scan dataset of the Kijkduin beach-dune system, The Netherlands. Scientific Data, 9 (1), pp. 191. doi: 10.1038/s41597-022-01291-9.
More information on the method of 4D objects-by-change can be found on this website: https://www.uni-heidelberg.de/4dobc
Find out more about the CoastScan project here: https://coastscan.citg.tudelft.nl/