A 3D terrain model from drone data turns a flat aerial photo into a surface you can measure, visualise, and analyse. It is built from the same photogrammetry process as an orthomosaic, but instead of a single top-down image, it produces a digital surface model or digital terrain model: a grid of elevation values that shows hills, valleys, buildings, and vegetation height. Understanding the difference between those models is the first step to using them well.

A digital surface model, or DSM, includes everything the drone sees from above: rooftops, trees, and other objects, as well as ground. A digital terrain model, or DTM, attempts to remove those objects and show bare ground elevation only. For flood modelling, earthworks, and contour mapping, a DTM is usually more useful, while a DSM is better for visibility analysis, building height, or tree canopy studies. Photogrammetry software can generate both, but a DTM from photos is less reliable under dense vegetation than one from LiDAR.

Flight planning affects terrain quality. Fly with high overlap, consistent altitude, and a camera angle that includes some oblique views if you need vertical faces, such as cliffs or building sides. Ground control points improve absolute accuracy, so the model's elevations match real-world heights rather than just relative shape. Without them, a model may look correct but be offset vertically by metres.

Once you have a terrain model, open it in QGIS and create a hillshade for visualisation, contours for a topographic map, and slope and aspect layers for analysis. Hillshade makes terrain readable at a glance, contours turn elevation into a printable map, and slope highlights steep areas relevant to erosion, landslides, or construction suitability. Together, they turn a grid of numbers into decisions.

Check accuracy before relying on a model. Compare elevations at known points, look for artefacts such as spikes over water or trees, and note the model's resolution: a model built from 5-centimetre drone imagery will show far more detail than a 30-metre satellite model, but it will also show noise that a coarser model smooths away. Document resolution, date, and processing settings alongside the model.

Drone terrain models are used for quarry volume calculations, construction progress, agricultural drainage, archaeology, and environmental monitoring, among many other tasks. They are fast to produce, relatively low cost, and repeatable, so the same site can be modelled monthly to track change. Their limits are vegetation, water, and very steep or reflective surfaces, where photogrammetry struggles to find matching features.

In short, a 3D terrain model from drone data is a powerful, practical product when its accuracy and limits are understood. Fly carefully, process thoughtfully, check against known points, and you will have a terrain dataset that supports real decisions, not just a pretty 3D view.

Before using a terrain model for decisions, check it against independent data: a surveyed point, a known benchmark, or a second dataset. Compare elevations at several locations, not just one, and note the differences. If differences are large or systematic, revisit flight planning or ground control before trusting the model for measurements such as volumes or flood levels.

Also share accuracy information with users of your model. A terrain model without a stated accuracy, resolution, and date is easy to misuse, because readers may assume it is more precise than it is. Clear metadata protects both you and the people who rely on your work.

Treat every terrain model as a measurement with limits, not as a perfect replica of the ground, and it will serve you well for planning, analysis, and communication.