Vector data cubes for features evolving in space and time
The amount of geospatial data generated, in particular from segmentation techniques applied to Earth observation (EO) data, is rapidly increasing. This, in combination with the rising popularity of EO data cubes for time series analysis, results in a need to adequately structure, represent and furth...
| Published in: | AGILE: GIScience Series |
|---|---|
| Main Authors: | , , |
| Format: | Text |
| Language: | English |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://doi.org/10.5194/agile-giss-5-16-2024 https://agile-giss.copernicus.org/articles/5/16/2024/ |
| Summary: | The amount of geospatial data generated, in particular from segmentation techniques applied to Earth observation (EO) data, is rapidly increasing. This, in combination with the rising popularity of EO data cubes for time series analysis, results in a need to adequately structure, represent and further analyse data coming from segmentation approaches. In this study, we explore the use of vector data cubes for the structuring and analysis of features that evolve in space and time with a particular focus on geomorphological features due to their high spatio-temporal variability. Vector data cubes are multi-dimensional data structures that often contain spatio-temporal data with n-dimensions, with a geometry as the minimum spatial dimension and time as the temporal dimension.We consider two vector data cube formats, i.e., array and tabular, and further extend their conceptualisation to contain features that evolve in space and time.We showcase our implementation for two geomorphological features, the Fagradalsfjall lava flow in Iceland and the Butangbunasi landslide and landslide-dammed lake in Taiwan. Finally, we discuss the potential and limitations of vector data cubes, regarding their technical implementation and application to geomorphology, and further outline the future research directions. |
|---|