DSM vs DTM vs DEM: What Is the Difference in Drone Mapping?

Drone mapping software can turn aerial imagery into much more than a photograph.

One of the most useful outputs is elevation data.

Instead of showing only what the site looks like from above, an elevation model describes how the height of the surface changes across the project area.

That information can support terrain analysis, drainage review, stockpile calculations, construction documentation, site planning, visualization, and many other mapping workflows.

The terminology can become confusing quickly.

Three acronyms appear constantly:

• DSM
• DTM
• DEM

They are related, but they do not always describe the same thing.

Understanding the difference is important because choosing the wrong elevation product can lead someone to interpret buildings, vegetation, equipment, or other surface objects as terrain.

What is a Digital Elevation Model?

Digital Elevation Model, usually shortened to DEM, is the broadest term of the three.

A DEM is a digital representation of elevation across an area.

The surface is commonly stored as a raster grid.

Each grid cell represents an elevation value.

Instead of each pixel representing a color like an aerial photograph, each pixel represents height.

That allows software to analyze how elevation changes throughout the site.

A DEM can be used to visualize:

• Slopes
• Ridges
• Depressions
• Drainage patterns
• Stockpiles
• Excavations
• Terrain changes
• Surface elevation

The complication is that DEM is sometimes used as a general category while other software or organizations use it to describe a specific elevation product.

That is why the terms DSM and DTM provide more useful detail.

They describe what type of surface the elevation model is intended to represent.

What is a Digital Surface Model?

A Digital Surface Model, commonly called a DSM, represents the elevation of visible surfaces.

The word surface is important.

If the drone sees the top of a building, the DSM may represent the roof.

If it sees the top of a tree canopy, the DSM may represent the vegetation.

If construction equipment is present, parts of that equipment may appear in the reconstructed surface.

A DSM can therefore include:

• Ground
• Buildings
• Roofs
• Vegetation
• Stockpiles
• Vehicles
• Construction equipment
• Trailers
• Material piles
• Other visible objects

This makes DSMs extremely useful for understanding the current condition of a site.

How is a DSM created from drone imagery?

Photogrammetry software analyzes overlapping photographs captured from different positions.

It identifies common features between images and reconstructs their positions in three dimensional space.

That reconstruction can produce a dense point cloud.

The point cloud can then be converted into an elevation surface.

When the resulting surface represents the visible top of everything captured, the product behaves as a DSM.

For example, imagine a construction site containing:

• A partially completed building
• Several soil piles
• Excavated trenches
• Material storage areas
• Temporary trailers
• Open ground

The DSM may represent all of those visible surfaces.

That can be extremely useful because the model captures the physical condition of the site at the time of the flight.

What is a Digital Terrain Model?

A Digital Terrain Model, commonly called a DTM, attempts to represent the underlying terrain.

Instead of preserving every visible surface object, the goal is to model the ground.

Objects such as:

• Buildings
• Trees
• Vehicles
• Equipment
• Trailers
• Temporary structures
• Other non ground features

may need to be removed or classified before the terrain model is created.

This makes DTMs particularly useful when the actual shape of the ground is more important than the complete visible surface.

DSM versus DTM

The simplest distinction is:

A DSM represents the top of visible surfaces.

A DTM attempts to represent the terrain beneath those surface objects.

Consider a tree.

A DSM created over that location may represent the top of the canopy.

A DTM attempts to represent the ground underneath the tree.

Consider a building.

A DSM may represent the roof elevation.

A DTM attempts to represent the terrain that the building occupies.

That difference can completely change the result of terrain analysis.

A practical example

Imagine a drone mapping mission over a land development project.

The site contains:

• Cleared ground
• Existing trees
• A construction trailer
• Several stockpiles
• Grading equipment
• A partially built structure

A DSM may contain elevation values for all of these objects.

If someone analyzes drainage directly from that surface, the building and stockpiles may influence the apparent flow paths.

A DTM attempts to remove those non terrain features and represent the underlying ground.

For grading or terrain analysis, that difference matters.

For construction progress documentation, however, the DSM may be more useful because the visible objects are part of what the project team wants to document.

Neither product is automatically better.

They answer different questions.

DSM versus DEM

A DSM is a type of elevation model.

DEM is often used as the broader category.

In many workflows, someone may casually call a DSM a DEM because it is a digital elevation model.

This is one reason file names alone are not always enough to understand what a dataset represents.

Documentation should state whether the elevation surface represents:

• Visible surfaces
• Bare terrain
• Another processed surface

Users should not have to guess.

DTM versus DEM

A DTM can also be considered a type of digital elevation model.

The key distinction is that a DTM is intended to describe terrain rather than the top of every visible object.

Different software platforms and geospatial organizations sometimes use these terms differently.

That variation is important.

Professionals should pay attention to methodology rather than relying only on the acronym.

Ask what the surface actually contains.

That is more useful than assuming every product labeled DTM or DEM was created the same way.

Why photogrammetry cannot always see bare earth

One of the most important limitations of drone photogrammetry is visibility.

Photogrammetry reconstructs surfaces using what appears in photographs.

If the camera cannot see the ground, the software does not magically know exactly where the ground is.

Dense vegetation is the clearest example.

If a forest canopy completely blocks the ground, aerial photographs primarily contain information about the canopy.

Software may attempt to classify or interpolate terrain, but that is not equivalent to directly measuring the hidden surface.

This matters when someone expects a drone photogrammetry DTM to represent accurate bare earth underneath dense vegetation.

The source data must support the intended result.

Why LiDAR can behave differently

LiDAR and photogrammetry create spatial data differently.

Photogrammetry reconstructs geometry by comparing visible features between photographs.

LiDAR directly measures ranges using emitted laser pulses.

Some LiDAR systems can produce returns from gaps in vegetation that allow more ground information to be captured beneath a canopy.

That can make LiDAR advantageous for certain terrain mapping applications.

It does not mean LiDAR automatically creates a perfect terrain model.

Point classification, capture density, vegetation conditions, positioning quality, and processing methodology still matter.

The important distinction is that the sensors observe the environment differently.

Reality-capture teams often choose between photogrammetry and LiDAR based on what the cameras or scanner can actually observe, not on the filename of the finished surface.

DSM applications

Digital Surface Models can be valuable for many drone mapping workflows.

Construction documentation

A DSM can preserve the visible condition of a construction site at a specific moment.

It can help teams understand:

• Stockpile locations
• Excavation conditions
• Building progress
• Temporary site features
• Material placement
• Large surface changes

Repeated captures can create a useful historical record. Drone mapping for construction covers how those dated surfaces support project tracking.

Construction teams often need the visible site, not a stripped terrain surface, when the goal is documenting what was actually there.

Stockpile analysis

Because a DSM represents visible surfaces, it can be useful for calculating the shape and volume of material piles when the mapping workflow and reference surface are appropriate.

The accuracy of any volume calculation still depends on capture quality, positioning, processing, surface definition, and the base surface used for comparison.

Site visualization

Colorized elevation maps can make site topography easier to understand.

Elevated structures, depressions, material piles, and terrain changes can become much easier to identify than they are in a normal aerial image.

Surface change detection

When repeat mapping missions are positioned consistently, elevation surfaces can help visualize changes between capture dates.

This may be useful for:

• Grading progress
• Excavation
• Fill placement
• Material movement
• Site development

The datasets need compatible coordinate systems and appropriate accuracy for meaningful comparison.

DTM applications

DTMs become more useful when terrain itself is the focus.

Grading analysis

A terrain model can help represent how the ground changes across a project.

This can support comparisons between existing conditions and proposed designs when used within an appropriate professional workflow.

Drainage understanding

Terrain elevation influences how water moves across a site.

A DTM can help visualize:

• High points
• Low points
• Slopes
• Potential flow paths
• Drainage patterns

Hydrologic or engineering decisions should still use data and methodology appropriate for the required level of accuracy. Surveying and mapping teams treat that as a methodology question, not a filename question.

Land development

Terrain information is fundamental to many land development workflows.

A DTM can provide useful spatial context for understanding the existing shape of a site.

Contour generation

Elevation models can be used to generate contour lines.

Contour quality depends directly on the quality and intended use of the elevation source.

Generating contours does not improve the accuracy of the source surface.

Contours are another representation of the underlying elevation data.

Elevation models and orthomosaics

An orthomosaic and an elevation model often come from the same photogrammetry project, but they contain different information.

The orthomosaic represents visual appearance.

The elevation model represents height.

An orthomosaic pixel usually stores image color information.

An elevation raster pixel stores an elevation value.

Viewed together, they provide much more context than either product alone.

For example, an orthomosaic may clearly show a soil pile.

The DSM can show its height and shape.

The point cloud can provide the underlying three dimensional points.

A 3D model can make the feature easier to visually interpret.

These products are related but not interchangeable.

Elevation models and point clouds

A point cloud represents many individual three dimensional points.

An elevation raster simplifies that three dimensional information into a grid.

Each raster cell contains an elevation value derived from the underlying surface.

Point clouds preserve more detailed three dimensional geometry.

Elevation rasters are often easier to analyze using GIS tools designed for terrain and raster calculations.

The appropriate representation depends on the task.

For a deeper explanation of three dimensional point data, see the SpearAtlas guide explaining what a point cloud is.

GeoTIFF elevation models

DSM, DTM, and other elevation rasters are commonly exported as GeoTIFF files.

This sometimes creates confusion because orthomosaics can also use GeoTIFF.

Both files may end with the same TIFF extension.

They contain different types of information.

An orthomosaic GeoTIFF normally stores imagery.

An elevation GeoTIFF stores elevation values.

That distinction is important when organizing and delivering mapping projects.

File type alone does not describe the entire purpose of a deliverable. Drone mapping deliverables explained covers how those files sit next to point clouds, meshes, and reports.

Hillshade

A hillshade is a visualization derived from elevation data.

Software simulates how light would illuminate the terrain from a chosen direction.

This creates highlights and shadows that make subtle surface changes easier to see.

Hillshades are useful for visual interpretation.

They are not separate measurements of elevation.

They are visual products derived from the elevation model.

Contours

Contour lines connect locations with the same elevation.

They provide a familiar way of representing terrain on a two dimensional map.

Contours can be generated from DSMs or DTMs, but the result will reflect whichever surface was used.

Contours generated from a DSM may follow buildings, vegetation, or stockpiles.

Contours generated from an appropriately classified DTM should more closely describe the terrain.

Always know which elevation source generated the contours.

Colorized elevation maps

Elevation surfaces can also be displayed using colors.

Different elevation ranges receive different colors, making high and low areas easier to identify visually.

These maps can be very effective for:

• Construction communication
• Grading visualization
• Terrain interpretation
• Client presentations
• Site comparisons

The colors are a visualization layer.

The underlying elevation values remain the important data.

Resolution versus accuracy

A very detailed elevation raster is not automatically highly accurate.

Raster resolution describes the spacing or size of elevation cells.

Accuracy describes how closely those elevation values correspond to reality.

A model can have very small cells and still contain positional or elevation error.

This distinction mirrors the difference between GSD and accuracy in photogrammetry. Drone photogrammetry accuracy explained covers why a sharp product is not automatically in the right place.

High resolution does not eliminate the need for proper positioning, control, checkpoints, processing, and quality assurance when accuracy matters.

Ground control and elevation models

Ground control can help constrain a photogrammetric reconstruction to known coordinates.

Checkpoints can help independently evaluate how closely the finished project agrees with known positions.

RTK or PPK can improve image positioning.

These tools influence the quality of the broader photogrammetry project from which elevation models are generated.

However, accurate camera positioning does not automatically make every reconstructed surface correct.

Vegetation, reflections, poor image coverage, weak geometry, moving objects, and processing artifacts can still affect the surface.

Elevation quality depends on the entire workflow.

Vertical accuracy matters

Elevation products deserve particular attention to vertical accuracy.

A small vertical error may have little consequence for general visualization.

The same error can matter much more when calculating:

• Volumes
• Cut and fill
• Drainage
• Slope
• Grade
• Elevation differences
• Construction quantities

The required accuracy should match the decision being made from the data.

A map created for visual documentation should not automatically be treated as appropriate for engineering or survey decisions.

What should be delivered with an elevation model?

A professional elevation deliverable should provide enough context for the recipient to understand what the file represents.

Depending on the project, useful information may include:

• Surface type
• Capture date
• Coordinate reference system
• Vertical reference
• Raster resolution
• Positioning method
• Control methodology
• Checkpoint information
• Processing methodology
• Known limitations
• Intended use
• File format

This becomes especially important when multiple elevation products exist in the same project.

A folder containing files called DSM, DTM, DEM, contours, and hillshade can quickly become confusing without context.

Naming elevation files clearly

Good file naming reduces mistakes.

Instead of:

surface.tif

Consider names that communicate what the file contains.

For example:

• site name DSM date.tif
• site name DTM date.tif
• site name hillshade date.tif
• site name contours date.dwg

The exact naming convention matters less than consistency.

A recipient should be able to identify the surface type without opening every file.

Organizing elevation products with the rest of the mapping project

Elevation models rarely exist alone.

The same drone mission may also produce:

• Orthomosaic
• Point cloud
• 3D model
• Source photographs
• Reports
• CAD files
• Control data
• Construction documentation
• Supporting spreadsheets

Keeping these outputs associated with the same project makes the mapping record easier to understand.

A project manager may primarily review the orthomosaic.

A technical user may download the elevation GeoTIFF.

Another team member may need the point cloud.

Someone else may only need a PDF report.

The delivery workflow should preserve the native data while making the relationship between those deliverables clear.

DSM, DTM, and DEM comparison

The practical difference can be summarized simply.

Always verify what the dataset actually represents.

Which one should you use?

Choose the elevation model based on the question the project needs to answer.

If the goal is:

What does the visible site surface look like right now?

A DSM may be appropriate.

If the goal is:

What is the shape of the underlying terrain?

A DTM may be more appropriate.

If a file is simply labeled DEM:

Determine how the surface was generated and what it represents before using it.

The acronym should never replace understanding the methodology.

The correct elevation model starts with the project objective

Drone mapping deliverables should be selected before the aircraft leaves the ground.

If terrain analysis is required, the capture and processing workflow should be designed around that requirement.

If construction documentation is the priority, preserving the visible surface may be more useful.

If engineering or surveying decisions depend on elevation values, appropriate professional standards, control, verification, and methodology become essential.

The strongest mapping workflow begins with a simple question:

What decision will someone make from this data?

That answer determines whether a DSM, DTM, or another elevation product is appropriate.

Understanding the difference between these surfaces prevents one of the most common mistakes in geospatial data use: assuming every elevation file represents the ground.