Refining satellite laser altimetry geolocation through full-waveform radiative transfer modeling and matching
Advances in geolocation accuracy for spaceborne laser altimetry data are critical for numerous applications. Existing methods often address the problem on a global scale, primarily relying on ground returns and overlooking the effects of vegetation canopy, which can lead to inaccuracies, especially...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-12-01
|
| Series: | Science of Remote Sensing |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2666017225000549 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Advances in geolocation accuracy for spaceborne laser altimetry data are critical for numerous applications. Existing methods often address the problem on a global scale, primarily relying on ground returns and overlooking the effects of vegetation canopy, which can lead to inaccuracies, especially when integrating these data with other georeferenced datasets. We developed a novel, site-focused approach to evaluate and correct geolocation errors in full-waveform spaceborne laser altimeter data using simulated data from the 3D discrete anisotropic radiative transfer (DART) model. Our analysis operates at the scale of a LiDAR footprint, employing waveform matching that effectively accounts for both vegetation and ground returns across various forest types. We used DART to (i) prepare realistic 3D vegetation scenes reconstructed from dense (>20pt/m2), small footprint airborne LiDAR data, and (ii) simulate Global Ecosystem Dynamics Investigation (GEDI) waveforms within these scenes. Our approach determines the “true” GEDI footprint positions by maximizing the similarity between simulated and observed GEDI full-waveforms within a local search area. We evaluated this method across various sites with different forest canopy types, finding strong correlations between simulated and observed GEDI waveforms (r2∈[0.94,0.99]) and root mean square errors (RMSE) ∈[0.14,0.63]. Random GEDI geolocation errors ranged from 5-10 m, and systematic errors were less than 8 m, within the GEDI product specification. Waveform matching was most successful for complex waveforms over heterogeneous, open canopies, and less effective for homogeneous, closed canopies over flat terrain. Our approach offers improved performance, with lower RMSE and higher correlation, over existing ALS-basedmethods. These advances support spaceborne data fusion through precise integration of vertical vegetation structure profiles from laser altimetry with horizontal vegetation structure of canopy and ground surface topography. |
|---|---|
| ISSN: | 2666-0172 |