Written by Ceezaer Drone Services
Official publication of ceezaer.com — Veteran-Owned · FAA Part 107 · Round Rock, TX
Official publication of ceezaer.com — Veteran-Owned · FAA Part 107 · Round Rock, TX
You hired a drone crew, and two days later a folder lands in your inbox full of files called an orthomosaic, a DSM, a DTM, and a point cloud. Each one answers a different question about your site, and paying for the wrong one wastes money and time. This guide breaks down all four drone deliverables in plain language, with the accuracy figures, file formats, and real use cases Central Texas contractors and developers actually need.
Every drone mapping flight produces raw images, but raw images are not the product. The value is in the processed drone deliverables built from those images: orthomosaics, digital surface models, digital terrain models, and point clouds. Each is a different way of representing the same ground.
When a drone flies a mapping mission, it captures dozens or hundreds of overlapping photos, usually with at least 70 percent front and side overlap so photogrammetry software can find matching points across frames. That software then reconstructs the site in three dimensions and outputs several distinct files. Understanding which file answers your question is the difference between a report you can act on and a folder you never open.
The four core drone deliverables fall into two groups. Orthomosaics are flat, map-accurate images you measure across. DSMs, DTMs, and point clouds carry elevation data, so you measure heights, slopes, and volumes. Most commercial projects need a combination, and knowing the distinction keeps your scope and your invoice honest.
A single stitched, geometrically corrected aerial image. Measure distances and areas as if it were a paper map.
Digital Surface Model. Elevation of everything on the ground: buildings, trees, equipment, and stockpiles included.
Digital Terrain Model. Bare-earth elevation with structures and vegetation removed to show the true ground surface.
Millions of individual 3D points, each with position and color. The raw geometric backbone for detailed analysis.
The most requested drone deliverable is also the easiest to understand. An orthomosaic looks like a photo but behaves like a map.
A single aerial photo has distortion. Objects lean, scale changes from the center of the frame to the edges, and the terrain is not accounted for. An orthomosaic fixes all of this. The software removes lens distortion, corrects for camera tilt, and adjusts for elevation so every pixel sits in its true horizontal position. The result is a seamless, top-down image where a measurement in one corner is just as accurate as a measurement in the other.
Resolution is described as ground sample distance, or GSD, the real-world size each pixel represents. A typical mapping flight at 200 to 300 feet produces a GSD of roughly 1 to 2 centimeters per pixel, sharp enough to read pavement cracks, count parking stalls, or trace a property boundary. Delivered files are usually GeoTIFF for engineering use or a compressed JPEG or PDF for client review.
Contractors use orthomosaics to measure disturbed area for stormwater compliance, verify that grading matches the plan, and document site conditions on a specific date. Real estate and land developers use them to show acreage, road frontage, and layout in a single frame. Because the file is georeferenced, it drops directly into AutoCAD, Civil 3D, or GIS platforms without manual scaling.
DSMs and DTMs both store elevation, but they answer opposite questions. Confusing the two is the most common and most expensive mistake in ordering drone deliverables.
A Digital Surface Model records the height of the topmost surface at every point. That means the DSM includes the roof of a building, the crown of a tree, a parked excavator, and the peak of a gravel stockpile. It is a raster grid where each cell holds an elevation value, usually delivered as a GeoTIFF that reads as a grayscale or color-shaded height map. If you want to know how tall something is or calculate the volume of material sitting on the ground, the DSM is your file.
A Digital Terrain Model strips all of that away. Through automated and manual filtering, structures, vehicles, and vegetation are removed, leaving only the bare earth. The DTM shows what the ground actually does underneath everything on it, which is exactly what a civil engineer needs for cut and fill calculations, drainage design, and slope analysis. Producing a clean DTM takes more processing time than a DSM because the classification and cleanup step is labor intensive.
Overlapping images are flown at consistent altitude with ground control points placed and surveyed for accuracy.
Photogrammetry software builds a dense 3D model and generates a DSM from the topmost visible surface.
Points are sorted into ground versus non-ground. Buildings, vegetation, and equipment are filtered out.
The cleaned bare-earth grid becomes the DTM, ready for engineering, drainage, and earthwork use.
A practical example from Central Texas earthwork: to calculate how much dirt a contractor moved this month, you compare two DSMs from different dates over the stockpile area. But to design the finished grade and drainage for that same pad, the engineer works from the DTM. Both come from the same flight; the deliverable you request drives the processing effort and price.
If the orthomosaic and elevation models are finished products, the point cloud is the source material that everything else is built from.
A point cloud is exactly what it sounds like: millions of individual points floating in three-dimensional space, each with an X, Y, and Z coordinate and, in photogrammetry, an RGB color value pulled from the images. A single acre captured at high density can hold well over 200 million points. Fly over the point cloud in viewing software and it renders as a recognizable, full-color 3D replica of the site that you can rotate, slice, and measure from any angle.
The standard file format is LAS or its compressed cousin LAZ, which most engineering and survey platforms read natively. Point clouds are the deliverable when a project needs cross sections, detailed volume calculations across irregular shapes, clash detection against a BIM model, or as-built verification against design coordinates. They are also the input from which orthomosaics, DSMs, and DTMs are generated, which is why a well-flown mission can produce all four deliverables from one visit.
Point clouds can come from photogrammetry or from LiDAR sensors. Photogrammetry point clouds carry true color and are cost effective for open sites, while LiDAR penetrates vegetation better and performs in low light. For most Central Texas commercial sites without heavy tree cover, a photogrammetric point cloud delivers survey-grade results at a lower cost. The tradeoff between the two methods deserves its own comparison, but the deliverable format itself is identical.
The right drone deliverable depends entirely on the decision you are trying to make. Order the file that answers your question, not the longest possible list.
Start with the outcome. Do you need to measure across a flat plane, such as area, distance, or a property layout? An orthomosaic is enough. Do you need to know how much material is present or how tall something is? You need a DSM, and for volume work, two DSMs from different dates. Are you designing grades, drainage, or earthwork quantities against a plan? Request a DTM. Do you need to slice cross sections, verify as-builts, or feed a BIM workflow? The point cloud is the deliverable.
Accuracy is the other lever. Adding surveyed ground control points brings horizontal accuracy into the 1 to 3 centimeter range, which is what engineers and surveyors expect for design work. Without ground control, a flight relying only on onboard GPS may drift to a foot or more, which is fine for marketing imagery but not for cut and fill. Always confirm whether your deliverable is tied to real-world control before it drives a construction decision.
A single well-planned flight can generate all four products, so the practical question is which ones you actually pay to have processed. At Ceezaer, mapping projects in the Austin metro are scoped around the decision the client needs to make, with AI-assisted reporting layered on top of the base deliverables and a standard 48-hour turnaround. The goal is a file you open and use, not a folder that sits unused because it answered the wrong question.
© 2026 Ceezaer™ Drone Services. This article was written and published by Ceezaer (ceezaer.com). All rights reserved — reproduction or republication without written permission is prohibited. Original URL: https://ceezaer.com/blog/drone-deliverables-orthomosaics-dsm-dtm-point-clouds
Shows what the deliverables in this guide actually cost on real Texas sites.
Explains the two methods behind the point clouds discussed here.
A deeper look at the flat, measurable deliverable covered above.