Ground truth, refined: The new photogrammetric Terrain Modeling in UltraMap
12 June 2025
With the release of UltraMap 6.6, we’ve taken a significant leap forward in the generation of Digital Terrain Models (DTMs). This latest update brings major improvements in accuracy, consistency, and usability – while preserving the trusted core methodology and user experience within the Ortho Pipeline module. The result? A globally consistent DTM at the push of a button – perfectly aligned and ready for quality check and integration into geospatial workflows anywhere in the world.
In this blog post, we’ll explore where and how the DTM has improved – and what that means for your projects.
A DTM represents the bare earth surface, with all vegetation, buildings, and above-ground objects removed. Unlike a Digital Surface Model (DSM), which captures everything visible from above, a DTM focuses exclusively on ground elevation. DTMs are foundational in a wide range of applications – from hydrological modeling and flood risk assessment to land development and orthophoto generation.
Terrain modeling with photogrammetry
Two major airborne technologies are used to create DTMs: photogrammetry and LiDAR. While both can produce terrain models, their underlying processes – and results – are quite different.
Photogrammetry reconstructs 3D surfaces by triangulating overlapping aerial images. The resulting point cloud includes everything visible – ground and non-ground alike. To create a DTM, the non-ground points (such as buildings, vegetation, vehicles) must be filtered out.
This is where the challenge lies: unlike LiDAR, photogrammetry doesn’t penetrate vegetation, making ground surface modeling more dependent on intelligent classification and filtering – since fewer or no ground points are visible beneath trees and in shaded areas.
At the same time, photogrammetry offers important advantages:
Lower acquisition costs compared to LiDAR, both in terms of equipment and flight operations
All data comes from the same flight, so the DTM as any other derived products are perfectly time-aligned with the images.
What’s new in UltraMap 6.6?
With version 6.6, we’ve made several major improvements to how UltraMap generates and refines DTMs:
1. Improved detection of ground and water
Our machine learning algorithms have been significantly enhanced to more accurately detect ground in complex environments, such as urban and vegetated areas. Non-ground features – including buildings, infrastructure, vegetation, moving objects, and temporary obstructions – are now filtered out with much greater confidence. Thanks to our dedicated water-handling approach, water surfaces are already well represented in the DSM and are carried over into the DTM. The result is a more accurate and consistent ground/non-ground classification.
Example of coast and harbor in Barcelona, Spain:
UltraMap leverages semantic information to intelligently handle non-ground areas.
Example of traffic infrastructure:
2. Enhanced Topographic Detail
Topographic features like mountain ridges, coastlines, and valleys are now preserved with far greater fidelity.
Our machine learning algorithms have been trained on a diverse range of global topographies.
Even in areas with dense vegetation, UltraMap achieves a better approximation of the actual ground level.
3. Cleaner infrastructure representation
Streets, embankments, and bridges are now rendered more cleanly and consistently. Users can still choose whether to include or exclude bridges in the DTM, depending on their use case – offering full flexibility within the pipeline. Additionally, users can now automatically select tiles where bridges have been detected. This feature simplifies quality control (QC) of bridges in the DTMOrtho and enables quick reprocessing of the DTM if a version without bridges is needed.
For DTMOrtho generation, bridges must be included in the underlying DTM to ensure that roads are correctly positioned. This prevents any displacement or misalignment in the ortho image texture. See the following DTM and ortho mosaic examples with and without bridges:
4. Improved visual representation
We’ve also enhanced the visual representation of the DTM, making it easier for users to assess terrain quality, spot inconsistencies, and validate results at a glance. By fine-tuning the noise levels across different areas of the DTM, we’ve achieved a more homogeneous appearance with clearly defined road structures and less noticeable transitions between interpolated and non-interpolated regions.
Fine-tuned noise leads to a more natural appearance of interpolated areas, making transitions smoother and less noticeable:
Why we invested in DTM – and how you can use it
One of the key strengths of a photogrammetric DTM is its seamless integration with high-resolution imagery. In UltraMap, DTMs have long served as the foundation for orthophoto generation – but with the latest improvements, they’ve evolved into a powerful product in their own right.
Thanks to a significantly enhanced baseline quality and effortless generation within the UltraMap workflow, DTMs now provide an excellent foundation for a wide range of geospatial applications. Built-in tools like the UltraMap Heightfield Editor allow users to make local corrections with ease. Edits to the DSM automatically update the derived DTM, with immediate visual feedback – streamlining the entire process.
This cost-effective and globally consistent output makes UltraMap DTMs a practical alternative wherever LiDAR-level accuracy isn’t required or traditional LiDAR methods are too expensive or logistically complex.
Typical use cases include:
- Building model creation, e.g., with RhinoTerrain – where roof structures can be combined with the DTM without the need for external data or LiDAR flights
- Hydrological modeling and watershed analysis
- Contour line extraction – the barer the terrain, the more accurately the DTM supports reliable results
- Terrain mapping during leaf-off season, where enhanced ground detection now ensures reliable DTM output
In short, the new UltraMap DTM is not just better – it’s more versatile, scalable, and accessible than ever before.
UltraMap 6.6 redefines what a photogrammetric DTM can achieve: more accurate, more consistent, and more accessible. Whether you’re generating ortho mosaics or working with terrain models directly, you’ll notice the improvements in every tile.
With smarter classification, enhanced topographic fidelity, and a more intuitive workflow, UltraMap’s DTM is now more versatile than ever.
Ready to test UltraMap’s enhanced DTM for yourself? Get access to high-resolution sample data processed with UltraMap 6.6.
Access sample data
